Effect of quenching in aqueous polyvinylpyrrolidone solutions on the microstructure and pitting corrosion resistance of AISI 1045 carbon steel

Carbon steels are widely used as infrastructure components owing to their high strength, ductility, and weldability. However, as corrosion degrades critical infrastructure components, exploring the relationship between the microstructure and corrosion resistance of carbon steel is of particular interest in engineering. To analyze this relationship, it is first essential to evaluate the corrosion resistance of each microstructure in a carbon steel specimen with a single chemical composition to eliminate the effects of alloying elements. In this study, we compared the pitting corrosion resistance of four microstructures (water‐quenched martensite, tempered martensite, bainite and degenerate pearlite, and ferrite–pearlite) of an AISI 1045 medium‐carbon steel. The pitting corrosion resistances of the microstructures in near‐neutral pH solutions containing NaCl were ordered as follows: (high) water‐quenched martensite > bainite and degenerate pearlite > tempered martensite > ferrite–pearlite (low). Elucidating the effect of the microstructure on the pitting corrosion resistance may contribute to the development of corrosion‐resistant carbon steels with excellent mechanical properties.

Carbon steels are generally used under atmospheric conditions, and long-term exposure to corrosive atmospheric conditions often results in the development of rust on the entire surface. Therefore, corrosion appears uniform under macroscopic observations. However, it is well known that corrosion initially occurs as pitting, [2][3][4][5][6][7][8] and the pitting-corrosion behavior of carbon steels in near-neutral pH solutions with chloride ions has been widely investigated. [9][10][11][12][13][14][15][16][17] For example, Katiyar et al. investigated and compared the corrosion behavior of five microstructures (pearlite, bainite, spheroidized, oilquenched martensite, and tempered martensite) of high-carbon steel. [15] Moreover, Kadowaki et al. analyzed the relationship between the microstructures and pitting corrosion resistance of AISI 1045 steel, which is among the most widely used medium-carbon steels. [16,17] Consequently, they discovered that the pitting corrosion resistances of typical steel microstructures were ordered as follows: (high) water-quenched martensite > tempered martensite ≈ primary ferrite > pearlite (low). [17] However, the pitting corrosion resistance of bainite in AISI 1045 steel is yet to be determined. Recently, steels containing bainitic structures have been intensively developed; hence, there is an urgent need to investigate the pitting corrosion resistance of bainite to contribute to the development of steels with desired mechanical properties and high corrosion resistance.
Bainite is formed at temperatures between the pearlite and martensite ranges on the time-temperature transformation (TTT) diagram. Particularly, the C curves of pearlite and bainite overlap significantly in plain carbon steels; therefore, bainitic structures are generally obtained by the addition of alloying elements that shift these C curves. [18][19][20] This, however, complicates the analysis of the effects exerted by alloying elements and microstructures on the pitting corrosion resistance. Vieira et al. reported a novel method for producing bainite in AISI 1045 steel by quenching in aqueous polyvinylpyrrolidone (PVP) solutions. [21] They discovered that the microstructure of a specimen quenched in a 20 mass% PVP solution contained approximately 60 vol% of martensite and 40 vol% of bainite. Their method enabled comparisons of the pitting corrosion resistances of bainite and other microstructures in AISI 1045 steel specimens with a single chemical composition.
The objective of this study was to determine the pitting corrosion resistance of bainite in AISI 1045 carbon steel. For this, specimens composed of martensite, bainite, and ferrite-pearlite structures were prepared using water and PVP solutions as quenchants. The pitting corrosion resistance of each microstructure was compared using potentiodynamic polarization in nearneutral pH solutions containing chloride ions.

| Specimens
A commercial AISI 1045 carbon steel sheet was used in this study, and its chemical composition is detailed in Table 1. The steel sheet was cut into coupons with dimensions of approximately 25 × 15 × 3 mm. One set of coupons was austenitized at 880°C for 1 h under vacuum, after which it was quenched in stirred water to obtain a completely martensitic structure. The specimens prepared using this method are referred to as "waterquenched" specimens in this paper. To obtain a tempered martensitic structure, the water-quenched specimens were heat-treated at 600°C for 1 h and quenched in stirred water. The specimens prepared using this method are referred to as "WQ-tempered" specimens in this paper.
Another set of coupons was austenitized at 880°C for 1 h under vacuum and quenched in aqueous PVP solutions with concentrations of 20 and 30 mass%, which cooled the samples more gradually than water. [21] Note that the PVP solutions were not stirred; therefore, although the specimens cooled rapidly immediately after immersion, the cooling rate decreased owing to the viscosity of the solutions and the steam film formed around the specimens. The exact cooling rate for each quenchant was not determined in this study; however, the solutions with higher PVP concentrations were found to result in slower cooling rates. [21] The specimens quenched in 20 and 30 mass% PVP solutions are referred to as "20%PVP" and "30%PVP" specimens, respectively, in this paper.
After heat treatment, the surfaces of the specimens were ground using a milling machine to eliminate decarburized layers. Next, the surfaces were ground using a series of SiC papers up to a grit size of 1500 grit, and these were then polished with 6 and 1 µm diamond pastes. Finally, the specimens were rinsed with ethanol.
T A B L E 1 Chemical composition of the steel sheet (mass%). The XRD data were collected with a step width of 0.02°and a scan speed of 2°/min. The Kα2 peaks were stripped (stripping ratio Kα2/Kα1 = 0.5) from the scans using computer software. Field-emission scanning electron microscopy (FE-SEM) was used to capture images of the specimen surfaces after 3 vol% nital etching. Secondary electron images were recorded at an accelerating voltage of 20 kV.
To analyze the volume fraction of the phases formed in the microstructures, color etching was conducted using a picral (2 mg of dry picric acid in 50 ml of ethanol) and 10 mass% Na 2 S 2 O 5 solution. The volume fraction of the phases was calculated via digital image processing of the optical micrographs obtained by color etching.
The microstructures of the specimens were characterized using SEM equipped with an electron backscattered diffraction (EBSD) detector. The specimens were observed using scanning transmission electron microscopy (STEM). Notably, the specimens for the EBSD analysis and STEM observations were prepared using a twin-jet electropolishing system with a mixture of HClO 4 (10 vol%) and ethanol (90 vol%). Final polishing was performed via ion milling. The accelerating voltage for the EBSD analysis was set at 25 kV. Backscattered electron (BSE) images were obtained at an accelerating voltage of 10 kV. The accelerating voltage for the STEM observations was 200 kV.

| Evaluation of the pitting corrosion resistance
To evaluate the pitting corrosion resistance in chloride environments, potentiodynamic anodic polarization was conducted in boric-borate buffer solutions with NaCl addition (pH 8.0) at 25°C under naturally aerated conditions. The solutions were prepared by mixing NaCl with 0.35 M H 3 BO 3 or 75 mM Na 2 B 4 O 7 . The NaCl concentrations were 5 and 10 mM. The measurements were performed on electrode areas of ∼10 mm 2 using a three-electrode cell. The counter electrode was a Pt plate, and the reference electrode was an Ag/AgCl (3.33 M KCl) electrode (0.206 V vs. a standard hydrogen electrode at 25°C). Note that all potentials reported herein refer to the standard hydrogen electrode. The scan rate of the electrode potential was 20 mV/min.

| Microstructure characterization
The phases in the specimens were identified via XRD, and the results are presented in Figure 1. Several peaks corresponding to the ferrite phase were observed at 44.6°, 64.9°, and 82.2°for the 20%PVP and 30%PVP specimens. In the water-quenched specimen, these peaks slightly shifted toward lower angles, indicating the presence of a martensite phase. No peaks corresponding to austenite were observed; thus, we can conclude that no retained austenite was present in the specimens.
After nital etching, the surfaces of the specimens were observed via FE-SEM, and the results are depicted in Figure 2. In the water-quenched specimen (Figure 2a was observed with elongated lath structures and cementite phases along the lath boundaries, which are characteristic of bainite. [22,23] In addition, discrete distributions of fine cementite phases were observed, indicating the formation of degenerate pearlite. Notably, previous studies have reported that degenerate pearlite generally forms at slightly lower temperatures than lamellar pearlite. [22][23][24] In the 30%PVP specimen (Figure 2c), primary ferrite and lamellar pearlite phases were observed. The surfaces of the primary ferrite phases were smooth; however, the lamellar structure of the ferrite and cementite phases made the surfaces of pearlite appear rough. Differences in the lamellar directions indicated different regions of the pearlite structure (pearlite colonies). The interlamellar spacing of pearlite was approximately 0.2 µm. The 20%PVP specimen probably had a bainitic structure because its structure differed from the entirely martensitic and ferrite-pearlite structures of the water-quenched and 30%PVP specimens, respectively.
The EBSD analysis was used to characterize the microstructures and grain orientations of the specimens. Figure 3 depicts the corresponding image quality and inverse pole figure maps of the specimens. In the waterquenched specimen (Figure 3a,b), a typical martensitic lath structure could be observed, which is consistent with the FE-SEM results (Figure 2a). In the 20%PVP specimen (Figure 3c,d), partial lath-like grains could be observed. Figure 4 depicts a BSE image of the area indicated by the dashed lines in Figure 3c,d. The results indicate that carbides precipitate inside the lath-like grains, suggesting that the lath-like grains in the 20%PVP specimens are bainitic ferrite phases. Note that the solubility limit of carbon in ferrite is considerably lower than that in austenite; therefore, insoluble carbon precipitates as a cementite phase during bainite transformation. In the 30%PVP specimens (Figure 3e,f), relatively coarse grains and no lath-like structures could be observed. The EBSD analysis also revealed that the 30%PVP specimens had a ferrite-pearlite structure.
To characterize the microstructures in detail, the specimens were observed using STEM; the bright-field images are depicted in Figure 5. Dislocations were observed in both the water-quenched ( Figure 5a) and 20%PVP (Figure 5b) specimens; however, the dislocation density of the 30%PVP specimen was relatively low. Furthermore, no lamellar structures associated with the ferrite and cementite phases were observed in the 20% PVP specimen. Notably, bainitic ferrite has been reported to exhibit a high dislocation density. [23,25] Therefore, the STEM results corroborate our observation that the 20% PVP specimen contained bainite.
Based on these results, it can be concluded that the microstructure of the 20%PVP specimen consists of bainite and degenerate pearlite. To analyze the volume fraction of bainite and degenerate pearlite, color etching was conducted using a picral and 10 mass% Na 2 S 2 O 5 solution. [26,27] This step tinted the bainite, pearlite, | 727 martensite, and ferrite phases as blue, dark brown, light brown, and white, respectively. [27] Figure 6 presents an optical micrograph of the 20%PVP specimen after etching. Two main color regions can be observed; these are tinted blue and dark brown. White lath structures, which appeared to be bainitic ferrite phases, were observed in the blue regions. The dark-brown regions indicated the presence of degenerate pearlite. The volume fractions of the blue (bainite) and dark-brown (degenerate pearlite) regions were determined to be approximately 52 and 48 vol%, respectively.
Schematics of the estimated cooling curves and a TTT diagram for hypoeutectoid steel are presented in Figure 7. Note that the TTT diagrams for carbon steels include the C curves for pearlite and bainite transformations. For the AISI 1045 carbon steel, these C curves overlap across a wide temperature range. [28] Primary ferrite forms at the F s temperatures above the pearlite nose temperature. In the 20%PVP specimen, the primary ferrite phase could not be observed via optical microscopy or SEM, indicating that the specimen rapidly cooled close to the nose temperature immediately after immersion in the PVP solution. The PVP solution was not stirred; hence, the cooling rate decreased owing to the high viscosity of the PVP solution and the vaporized layer that formed around the specimen. Therefore, the specimen passed through the temperature ranges of both the pearlite and bainite transformations. The formation of degenerate pearlite was attributed to the fact that the transformation occurred at a lower temperature compared to that of lamellar pearlite. [22][23][24] Few martensite phases were observed in the 20%PVP specimen (see Figure S1); therefore, it is believed that the pearlite and bainite transformations were largely complete. In the 30%PVP specimen, the cooling rate was further reduced; hence, a structure comprising primary ferrite and lamellar pearlite was obtained.
Previously, Vieira et al. reported that the microstructure of an AISI 1045 steel specimen quenched in a 20% PVP solution contained approximately 60 vol% of martensite and 40 vol% of bainite. [21] They quenched the specimens in PVP solutions that were continuously agitated using a recirculation pump with an outlet speed of 0.7 m/s. By contrast, the PVP solutions considered in the present study were not stirred, which resulted in little martensite formation. The exact cooling rates of the specimens in the PVP solutions were not recorded in this study; however, the 20%PVP specimen appeared to cool more gradually than the water-quenched specimen, and bainite was produced successfully.

| Evaluation of the pitting corrosion resistance
The pitting corrosion resistance of each specimen was investigated based on potentiodynamic polarization in boric-borate buffer solutions containing NaCl (pH 8.0). The polarization curves of the WQ-tempered specimens were also recorded because martensitic steels are generally tempered to balance their strength and toughness. The polarization curves of the specimens are presented in Figure 8. First, the experiments were conducted with a NaCl concentration of 10 mM (Figure 8a). For each specimen, potentiodynamic polarization began at approximately 0 V, and cathodic currents flowed first. The anodic currents were measured after the cathodic currents reached zero, and the anodic current density gradually increased to approximately 1 × 10 -2 A/m 2 . This current density represents the passive F I G U R E 6 Optical micrograph of the 20%PVP specimen after etching in a picral and 10 mass% Na 2 S 2 O 5 solution. Bainite, pearlite, martensite, and ferrite are tinted with blue, dark brown, light brown, and white, respectively. [27] [Color figure can be viewed at wileyonlinelibrary.com] F I G U R E 7 Schematic depicting the estimated cooling curves and a time-temperature-transformation diagram for hypoeutectoid steel. state of steel in near-neutral pH solutions. [14,29,30] Current oscillations in the passive state indicate metastable pitting events, and the subsequent continuous increase in current density is often attributed to the initiation of stable pitting. In our analysis, the pitting potential of the water-quenched specimen (black curve) was approximately 0.05 V higher than that of the other specimens. However, no substantial differences were observed between the pitting potentials of the WQtempered, 20%PVP, and 30%PVP specimens (gray, red, and blue curves, respectively).
Next, the NaCl concentration was reduced to 5 mM to investigate the differences between the pitting corrosion resistances of the specimens (Figure 8b). For the waterquenched specimen (black curve), a large increase in the current was observed at 0.36 V. For the WQ-tempered specimen (gray curve), stable pitting began at 0.29 V. The pitting corrosion resistance of tempered martensite in the WQ-tempered specimen was lower than that of asquenched martensite in the water-quenched specimen. For the 20%PVP specimen (red curve), pitting corrosion occurred at 0.31 V. This observation suggests that the pitting corrosion resistance of the microstructure comprising bainite and degenerate pearlite was lower than that of martensite in the water-quenched specimen and slightly higher than that of tempered martensite in the WQ-tempered specimen. For the 30%PVP specimen (blue curve), stable pitting occurred at 0.21 V, indicating that the pitting corrosion resistance of the ferrite-pearlite microstructure was lower than those of the other microstructures.
Optical micrographs of the surfaces of the specimens after polarization are presented in Figure 9. Pits with a diameter of approximately 30 µm were observed in each specimen. However, owing to the growth of the pits, the precise locations of the pitting initiation sites in the microstructures could not be determined. Therefore, whether the pitting corrosion resistance of bainite or degenerate pearlite had a greater effect on the pitting potential of the 20%PVP specimen remains unclear. However, these microstructures presented superior pitting corrosion resistances compared to ferrite-pearlite and tempered martensite. As presented in Figure 9a,b, some pits were formed adjacent to nonmetallic inclusions. Notably, manganese sulfide inclusions (MnS) were present in the specimens used in this study (see Figure S2), and sulfide inclusions are known to be preferential sites for the initiation of pitting corrosion in carbon steels. [14,[31][32][33] Therefore, it is likely that the pits observed herein were initiated from the MnS inclusions in each microstructure.
The pitting potentials of each specimen with 5 and 10 mM NaCl are presented in Figure 8c. The polarization curves were recorded twice under each condition, and the average values of the pitting potentials were plotted with error bars indicating the maximum and minimum values. When the NaCl concentration was 5 mM, the pitting corrosion resistances of the microstructures were ordered as follows: (high) water-quenched martensite > bainite and degenerate pearlite > tempered martensite > ferrite-pearlite (low). This is consistent with the order followed by the interstitial-carbon concentration. Kadowaki et al. reported that the higher pitting corrosion resistance of waterquenched martensite could be attributed to the interstitial carbon and that the pitting potential decreased as the interstitial-carbon concentration decreased with tempering. [31] Moreover, they proposed that interstitial carbon reduces the active dissolution rate inside pits. [34] Li et al. reported that the bonds between interstitial carbon and metal atoms in stainless steels were 1.4-2.0 times stronger than the metal-metal bonds. [35] Note that in this study, it was impossible to measure the interstitial-carbon concentration within each phase. However, it seemed reasonable to assume that the increase in the pitting corrosion resistance could be attributed to the reduction in the metal dissolution rate resulting from the interstitial carbon in the local environment of the pits and that interstitial carbon inhibited the transition from metastable pitting to stable pit growth.
As illustrated in Figure 8c, the pitting corrosion resistance of the bainite and degenerate pearlite microstructure was comparable to that of the tempered martensite in the WQ-tempered specimen. The effect of bainite on the pitting corrosion resistance of carbon steels has long remained unclear, despite the fact that it has attracted attention owing to its mechanical properties. This is because alloying elements are generally added to carbon steels to obtain a bainitic structure, [18][19][20] which makes it impossible to distinguish between the effects of such alloying elements and the microstructure on the pitting corrosion resistance. This study suggests that the pitting corrosion resistances of the microstructures of AISI 1045 steel can be controlled by the concentration of interstitial carbon.
Bainite can be considered a structure consisting of fine carbide precipitates (cementite) in ferrite. [36] According to this study, bainite exhibited superior pitting corrosion resistance compared to ferritepearlite, as presented in Figure 8c. This may be attributed to the higher interstitial-carbon concentration in bainite compared to that in ferrite-pearlite. In addition, previous studies have reported that the corrosion resistance of cementite is higher than that of ferrite and that cementite inhibits pit growth in the early stages of pitting corrosion. [16,37] Therefore, the difference between the pitting corrosion resistances of bainite and ferrite-perlite may also be related to the precipitation morphology of cementite; however, the corresponding details remain unclear and should be explored in future research. In summary, this study determined the order followed by the pitting corrosion resistance of martensite, bainite, and ferrite-pearlite microstructures of AISI 1045 carbon steel with a single chemical composition.

| CONCLUSION
Specimens with different microstructures (water-quenched martensite, tempered martensite, bainite and degenerate pearlite, and ferrite-pearlite) were produced from the same AISI 1045 carbon steel sheet using PVP solutions as quenchants. The pitting corrosion resistance of each microstructure was evaluated via potentiodynamic polarization in boric-borate buffer solutions containing NaCl (pH 8.0). The results revealed that the pitting corrosion resistances of the microstructures were ordered as follows: (high) water-quenched martensite > bainite and degenerate pearlite > tempered martensite > ferrite-pearlite (low). We expect that this clear understanding of the relationship between the microstructure and pitting corrosion resistance hastens the development of carbon steels with desirable mechanical properties and high corrosion resistance.

AUTHOR CONTRIBUTIONS
Masashi Nishimoto performed the experiments, analyzed the results, and drafted the manuscript. Izumi Muto conceived and supervised the study and revised the manuscript. Takashi Doi and Kaori Kawano analyzed the chemical composition of the steel sheet and revised the manuscript. Yu Sugawara revised the manuscript.