Enhanced Corrosion Resistance of Additively Manufactured 316L Stainless Steel After Heat Treatment

The chemical composition of the SLM SS316L is listed in Table I. SLM SS316L samples were obtained from an EOS M290 machine (Germany), which was equipped with a 400 W Yb-fiber laser, at a scanning speed of 980 mm s⁻¹. The total build volume was 250 × 250 × 325 mm. The size of the focused beam was 100 μm. The porosity of the as-built SLM SS316L sample was approximately 0.9% (measured by the mercury intrusion porosimeter (MIP) method).

Three heat-treatment conditions were used. (1) HT500: the temperature was raised to 500 °C and maintained for 4 h, which can release some residual stress.²⁸ (2) HT950: the temperature was raised to 950 °C and maintained for 4 h. (3) HT1100: the temperature was raised to 1100 °C and maintained for 1 h. HT1100 was chosen to generate full recrystallization. All the heat treatment processes were performed in a vacuum environment, followed by furnace cooling.

Electrochemical tests were performed in a traditional three-electrode cell; a Pt plate was the counter electrode, the sample served as the working electrode, and a saturated calomel electrode (SCE) acted as the reference electrode. The sample for the electrochemical test was embedded in plastic tubes and enclosed with epoxy resin. The work area was a circle 10 mm in diameter. Subsequently, the sample was ground with 80–2000 grit SiC paper, polished with 0.5 μm diamond paste, cleaned with ethanol and deionized water, and finally dried in air.

A solution containing 0.5 M H₂SO₄ + 2 ppm HF was used to simulate the PEMFC cathode environment. All electrochemical tests were performed at 70 °C with air bubbling. The sample was first cathodically polarized at −0.6 V_SCE for 5 min. To reach a steady state, all samples were immersed in the electrolyte for 3600 s. Subsequently, potentiodynamic polarization tests were performed at a potential of −0.3 V below the open circuit potential to +1.4 V at a rate of 1 mV s⁻¹. The potentiostatic polarization tests were conducted at a potential of 0.6 V_SCE for 4 h. EIS measurements were performed after potentiostatic polarization at 0.6 V for 4 h in a 0.5 M H₂SO₄ + 2 ppm HF solution. The frequency range was from 100 kHz to 10 mHz, and the sinusoidal amplitude was 5 mV. The Mott-Schottky test was conducted at 1 kHz with a scan rate of 50 mV s⁻¹ from 0.6 V_SCE to −0.5 V_SCE. All the electrochemical tests were conducted at least three times under each condition.

XRD was performed with a Panalytical X'Pert Pro (operated with Cu Kα radiation at 40 kV and 30 mA) to obtain phase information from the SLM SS316L samples before and after heat treatment. The samples were scanned at a rate of 5°/min in the 2θ range of 20°–80°. EBSD measurements were performed to obtain the grain sizes, grain orientations, and dislocation distributions. For the EBSD measurements, the collection speed, scanning step, and acceleration voltage were 637.76 Hz, 3 μm, and 20 kV, respectively.

Microstructural features were observed using SEM (FEI NovaNano SEM, equipped with energy dispersive spectroscopy (EDS), operated at 15 kV) with secondary electron (SE) imaging. To visualize the surface features, the samples were etched in aqua regia (HCl:HNO₃ = 3:1). The evolution of cellular structures and dislocations during heat treatment was observed using TEM (Hitachi H-9500, operated at 300 kV). The preparation process for the TEM samples comprised the following steps. First, a 3 mm disk was cut off from the sheet via wire cutting. Then, the disk was mechanically ground to 60 μm thick and punched into a 3 mm diameter disk. Finally, the small disk was double-jet electropolished using a Struers Tenupol-5 at −30 °C.

The chemical composition of the passive film was analyzed using XPS (ESCALAB 250xi, Thermo Fisher) with an Al Kα X-ray source (hv = 1486.68 eV, source power = 150 W). Before XPS surface analysis, all samples were potentiostatically tested at 0.6 V_SCE for 4 h in a 0.5 M H₂SO₄ + 2 ppm HF solution (70 °C). XPS results were obtained by analyzing three areas on each SLM sample. The peaks in the XPS spectra were investigated using the XPSPEAK 4.1 software after calibration of the C 1 s peak (284.8 eV) and subtraction of the baseline.

XRD patterns of the as-received SLM SS316L and heat-treated samples are shown in Fig. 1. Only the (111), (200), and (222) peaks of single-phase austenite were found in the as-received sample. The primary growth orientation was along the (111) direction, which is consistent with a previous report.³⁰ After heat treatment, the peak intensity of the SLM samples decreased markedly, and the (222) grain orientation disappeared.

Zoom In Zoom Out Reset image size

EBSD images of the SLM SS316L samples are shown in Fig. 2. As shown in Fig. 2a, the as-received sample had irregular and spindly grains, which are attributed to the rapid solidification process (usually 10⁵−10⁸ K s⁻¹). The as-received and HT950 samples (Fig. 2c) had almost the same grain size and grain shape, but the HT1100 sample (Fig. 2e) exhibited regular polygonal grains due to recrystallization. As seen in Figs. 2b, 2d, and 2f, the kernel average misorientation figures of the as-received and HT950 samples indicated the presence of a high-density of dislocations distributed inside the grains. However, the dislocation density of the HT1100 sample clearly decreased, indicating that the dislocation structures were eliminated after heat treatment at 1100 °C. As summarized in Table II, the average grain sizes did not change significantly between the as-received and HT950 samples. Grain growth was noticeable after heat treatment at 1100 °C, which is ascribed to recrystallization.

Zoom In Zoom Out Reset image size

Figure 3 displays the surface morphologies of the SLM SS316L samples after the potentiodynamic polarization test in a simulated PEMFC environment. From Figs. 3a and 3b, it is evident that the corrosion surface of the HT500 sample was similar to that of the as-received sample. The corrosion grooves at the cellular substructures and melt-pool boundaries were notably clear. Meanwhile, intense corrosion attacks occurred in areas close to melt-pool boundaries. Previous studies reported that melt-pool boundaries contain some defects, such as residual stress, a heat-affected zone, and a nonmetallic unstable state.³¹ Shifeng et al.³² reported that the concentration of nonmetallic elements (C, O, Si) near melt-pool boundaries greatly decreased. Therefore, the melt-pool boundary may be composed of an inhomogeneous microstructure that is preferentially corroded.^31,32 After heat treatment at 950 and 1100 °C, these cellular substructures were not observed, as shown in Figs. 3c and 3d, respectively. The melt-pool boundaries disappeared, and only the corrosion grooves at the grain boundary were clear.²⁶

Zoom In Zoom Out Reset image size

Porosity has an important effect on the pitting corrosion resistance of SLM SS316L.^33,34 In this work, the porosity of the as-received SLM SS316L sample was approximately 0.9%. In addition, the pore size of the as-received SLM SS316L samples was small (no pores were observed in the EBSD images). Sander et al.³⁴ reported that SLM SS316L samples with different porosities had similar pitting corrosion resistances. Schaller et al.³⁵ confirmed that large pores (>50 μm) were the primary reason for the reduced corrosion resistance in sample compared to that of samples with small pores (<10 μm). Therefore, in this study, it is thought that porosity had a marginal effect on the corrosion resistance.

Figure 4 shows an SEM image and a TEM image of the as-received sample. Distinct subgrain structures were observed in the etched surface morphology of the as-received sample, and the size of the subgrain was approximately 0.5–1 μm. Meanwhile, subgrain boundaries protruded from the surface compared to the center of the subgrain, indicating a better etch resistance. As shown in Fig. 4b, subgrain boundaries were entangled by a high concentration of dislocations, and dislocation walls revealed distinct contrast.

Zoom In Zoom Out Reset image size

Figure 5 shows an SEM image and a TEM image of the HT950 sample. As shown in Fig. 5b, the migration and rearrangement of dislocations occurred. Only some dislocation lines and overlapped dislocations were observed. However, subgrain structures are also shown in Fig. 5a. After heat treatment at 950 °C, the subgrain structure began to disappear. Compared to the as-received sample, the subgrain boundaries became discontinuous and formed a homogeneously distributed island-like cellular trace microstructure.

Zoom In Zoom Out Reset image size

Figure 6 displays the potentiodynamic polarization curves of SLM SS316L at 70 °C in a 0.5 M H₂SO₄ + 2 ppm HF solution. As observed, all the samples exhibited large passive regions in the potential range from 100 to 800 mV_SCE. Generally, the passive current density is related to the stability of the passive film.³⁶ The passive current densities and corrosion potentials obtained from the potentiodynamic polarization curves are listed in Table III. The as-received and HT500 samples had similar passive current densities, while the HT950 and HT1100 samples showed considerably lower passive current densities. This indicates that heat treatment at 950 and 1100 °C can greatly improve the stability of the passive film. Kong et al.²⁷ also found that a SLM SS316L sample heat treated at 1050 °C displayed a lower passive current in simulated PEMFC environments, which is in good agreement with the results reported here.

Zoom In Zoom Out Reset image size

Potentiostatic polarization tests were conducted at 0.6 V_SCE for 4 h at 70 °C in a 0.5 M H₂SO₄ + 2 ppm HF solution; the results are shown in Fig. 7. For all the curves, the passive current density decayed rapidly at the beginning and then gradually maintained a relatively stable value, indicating the formation of a passive film. After all the passive current densities were stable, the order of the passive current densities was as follows: HT950 > HT1100 > HT500 > as-received.

Zoom In Zoom Out Reset image size

EIS was performed after the potentiostatic tests at 0.6 V for 4 h at 70 °C in a 0.5 M H₂SO₄ + 2 ppm HF solution. The Nyquist plots of the as-received and heat-treated SLM SS316L samples are presented in Fig. 8. The Nyquist plots are composed of a capacitive arc. To calculate the resistance of the passive film, a simple equivalent circuit was employed (inset in Fig. 8). For the equivalent circuit, R_S represents the solution resistance, CPE is the constant phase element, and R₁ corresponds to the charge transfer resistance. Table IV lists the results obtained from fitting the EIS results. R₁ of the as-received and HT500 samples exhibited a relatively low value, while the R₁ value clearly increased after the high-temperature heat treatments. Moreover, the HT950 sample showed the highest charge transfer resistance, indicating that it had the best corrosion resistance. This result agrees well with the results of the potentiodynamic polarization curves (Fig. 6) and potentiostatic polarization curves (Fig. 7).

Zoom In Zoom Out Reset image size

Figure 9 shows the Mott-Schottky results of the SLM SS316L samples after the potentiostatic test at 0.6 V_SCE for 4 h at 70 °C in a 0.5 M H₂SO₄ + 2 ppm HF solution. The space-charge capacitance (C_SC) of the passive film of an n-type semiconductor can be calculated using Eq. 1^37,38:

$$\begin{eqnarray}&&{C}^{-2}=\displaystyle \frac{2}{\varepsilon {\varepsilon }_{0}e{N}_{D}}\left(E-{E}_{FB}-\displaystyle \frac{KT}{\varepsilon }\right)\end{eqnarray} \tag{ 1 }$$

where ε is the dielectric constant of the passive film, ε₀ is the vacuum permittivity (8.854 × 10^–14 F cm⁻¹), e is the electron charge (1.602 × 10^–19 C), N_D is the donor density (cm⁻³), E is the applied potential (V_SCE), E_FB is the flat band potential (V_SCE), K is the Boltzmann constant (1.38 × 10^–23 J K⁻¹), and T is the absolute temperature (K). As shown in Fig. 9, all the curves had a positive slope in the potential range from 0 to 600 mV, indicating the formation of an n-type semiconducting passive film.^39,40 Based on these linear domains, the donor densities of the passive film were obtained, as shown in Fig. 10. There was a high donor density in the as-received sample, which is attributed to the rapid cooling rate of SLM components. After heat treatment at 950 °C and 1100 °C, the donor density decreased considerably, indicating the formation of a more stable passive film. Morshed-Behbahani et al.⁴¹ also confirmed that a stabilization heat treatment (900 °C) was helpful in decreasing the defect concentration by homogenizing the distribution of Cr. Interestingly, the HT500 sample showed the highest donor density. This may be caused by the reduction in compressive residual stress after stress-relieving annealing. Similar results were also reported by Cruz et al. ⁴², where the donor density of the passive film increased after heat treatment at 450 °C and 650 °C.

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

To analyze changes in the composition after the various heat treatments, XPS was performed after the potentiostatic tests at 0.6 V_SCE for 4 h. The passive film primarily contained the oxides and hydroxides of Fe and Cr, as shown in Fig. 11. The high solution spectra of Fe 2p_3/2 were decomposed into three peaks: metallic Fe (707.3 eV), Fe₃O₄ (708.2 eV), and FeOOH (711.8 eV). There was a distinct metallic Fe peak, which is attributed to the thin passive film on SLM SS316L. The Cr 2p_2/3 ionization was deconvolved into metallic Cr (574.5 eV), Cr₂O₃ (576.1 eV), and Cr(OH)₃ (577.3 eV).

Zoom In Zoom Out Reset image size

Based on the fitted peak area, the cationic fractions of Cr and Fe and the Cr^ox+hy/Fe^ox+hy ratio are shown in Fig. 12. For all the samples, iron oxide was the primary component of the passive film. It is worth noting that the cationic fractions of the as-received, HT500, and HT950 samples were very similar. However, the Cr^ox+hy/Fe^ox+hy ratio of the HT1100 sample increased considerably. Cr(OH)₃ was enriched in the passive film after heat treatment at 1100 °C, while only a slight increase in Cr₂O₃ content was observed. This may be caused by the partial elimination of the subgrain structures after recrystallization annealing. Fine-grained microstructures can reduce the diffusion channels of Cr to the surface.⁴³

Zoom In Zoom Out Reset image size

In general, the pitting corrosion resistance of SLM SS316L is higher than that of wrought SS316L, because the rapid solidification process of SLM SS316L avoids the generation of inclusions (typically MnS). MnS inclusions may act as pit initiation sites and accelerate corrosion. According to a previous report²⁹, MnS inclusions appeared after heat treatment above 1000 °C and led to a decrease in pitting resistance. However, Kong et al. ⁴⁴ reported that no MnS inclusions were found after heat treatment at 1200 °C. This may be due to the low Mn concentration in SLM SS316L (0.42%), while the Mn concentration in other reports was above 1%. According to the previous research of our group²⁵, Si-rich and O-rich inclusions were found after heat treatment at 1100 °C, but no MnS inclusions appeared (probably due to the low Mn concentration (0.23%)). In addition, the homogenous distribution of Cr around the oxide inclusions may avoid the reduction in local corrosion resistance.²³ It is reasonable that inclusions have a certain degree of influence on corrosion resistance, but are not the primary factor.

Numerous studies have found that there are higher concentrations of Cr, Ni, and Mo at subgrain boundaries compared to the center of subgrains.^23,34,45 The increase in alloying elements helps subgrain boundaries resist the attack of aggressive ions.⁴⁶ This is beneficial to the corrosion resistance of subgrain boundaries. Kong et al. ⁴⁷ reported that the Volta potential at the subgrain boundary was higher than that at the subgrain interior owing to the higher Cr and Mo concentrations at the boundary. Therefore, subgrain boundaries appear to act as physical barriers to resist corrosion attacks. It can be seen that fine subgrain structures play an important role in improving corrosion resistance.⁴³ Revilla et al.⁴⁶ suggested that SLM SS316L samples showed larger passivity compared to a LMD (laser metal deposition) sample in a 3.5 wt% NaCl solution, which they attributed to the finer subgrain network in the SLM SS316L samples. In summary, the effect of the subgrain structure on the passive film properties should be considered.

For the HT500 samples, although the annihilation of dislocations probably occurred, the microstructures (Fig. 3b) and passive film composition (Fig. 12) of the HT500 sample was similar to that of the as-received sample. It appears that the elimination of residual stress had a marginal effect on the corrosion resistance, which is consistent with previous research.^27,42

After heat treatment at 950 °C, it can clearly be seen that dislocations were eliminated and rearranged (Fig. 5b) along with the partial annihilation of subgrain boundaries. However, it was reported that subgrain structures composed solely of dislocations (without Mo and Cr segregation) were found in additively manufactured 316L.⁴⁸ Therefore, both subgrain structures (with Cr and Mo segregation) and pure dislocation cellular structures coexist in additively manufactured 316L. For pure dislocation structures, it is well known that an increase in dislocation density reduces corrosion resistance.⁴⁹ The reduction in the local equilibrium potential in the dislocation-rich region promotes the dissolution of the anodic metal.⁵⁰ After heat treatment at 950 °C, some of the dislocations were eliminated, resulting in the improved corrosion resistance of HT950. However, the microstructural variation (island-like cellular trace structures in HT950) of the subgrain structures during heat treatment is a direct result of the annihilation of dislocations and rearrangement of subgrain boundaries. According to a previous report,²⁹ the alloying element segregation in subgrain structures showed good stability during heat treatment. This was caused by the distinct subgrain structures of SLM SS316L. During heat treatment at 950 °C, the pure dislocation subgrain boundaries quickly disappeared, while the subgrain boundaries with alloying element segregation were partially eliminated. This partial and incomplete homogenization of composition caused the formation of island-like cellular trace structures. However, it is difficult to obtain direct evidence (composition distribution of subgrains) owing to the limited instrument accuracy. However, from the EIS and passive current density results, the homogenization of the subgrain composition appears to be beneficial to the passivity of SLM SS316L. In general, high-temperature heat treatments (900 °C–1000 °C) improved the passive film properties of SLM SS316L.^29,51 This provides a new idea for optimizing the corrosion resistance of SLM SS316L.

Due to complete recrystallization, the dislocation density of the HT1100 sample was markedly reduced compared to the as-received and HT950 samples (Figs. 2b, 2d, and 2f). Dislocation structures (subgrains) were partially or almost completely eliminated during heat treatment at 1100 °C, and the grain size increased. The XPS results demonstrated that heat treatment at 1100 °C promoted the formation of Cr(OH)₃. Cr₂O₃ films have better corrosion resistance than Cr(OH)₃ films, and Cr(OH)₃ is formed by consuming Cr₂O_3.^52,53 The lower Cr₂O₃/Cr(OH)₃ ratio of the HT1100 samples compared to that of the HT950 samples resulted in a decrease in the passive film property after heat treatment at 1100 °C. However, the chemical segregation caused by recrystallization must be considered. Previous work from our group demonstrated that many nanoscale inclusions (enrichment of Si and O) are present in grains after heat treatment at 1100 °C. These inclusions may promote local corrosion.