Introduction

Austenitic 316L stainless (316L) has a wide range of industrial applications ranging from kitchenware to nuclear, and aerospace industries1. However, the production of complex structures of 316L by conventional manufacturing technologies entails high production time and cost and deteriorates the durability in many instances2. Additive manufacturing techniques like selective laser melting (SLM) are suitable for producing complex structures and can revolutionize the manufacturing and application of materials. SLM is a powder bed fusion AM technique and has earned massive research attention owing to its ability to manufacture complex structures in a single-step operation employing computer-aided design. Nominally, SLM involves selective melting of powders using a high-energy laser beam followed by fusing metallic powder at a solidification rate of ~(104–107) K/s3.

Successful production of 316L components manufactured by SLM has been reported extensively. Superior and reproducible corrosion properties of SLM printed components are crucial to extend their engineering applications. However, the corrosion behavior of SLM-316L SS reported in the literature is inconsistent. Some studies4,5,6 showed superior corrosion resistance of the SLM-316L than wrought-316L, whereas others7,8,9,10 showed inferior properties. The inconsistency in the corrosion performance has been attributed to the manufacturing defects (like porosity) and microstructural heterogeneities6,11,12. Post-processing is performed on the SLM-316L components to eliminate the manufacturing defects and obtain improved and consistent properties. However, the post-processing is time-consuming, expensive, causes material wastage, and sometimes does not result in the expected outcomes5. Deterioration in corrosion performance of SLM-316L has been reported after post-processing using hot-isostatic pressing (HIP) and heat treatments at elevated temperatures (900–1200 °C)5,13,14,15. One of the approaches to overcome the issues could be improving the corrosion resistance of the as-printed components. The strategy of feedstock modification can enhance the corrosion resistance, compensating the deleterious influence of the processing defects and microstructural heterogeneities, and the need for post-processing could be circumvented10,16,17,18,19,20 These modified composite feedstocks can be produced by (1) mechanical alloying of a suitable additive and feedstock powder17,20,21,22,23, (2) deposition/coating of additive on the surface of the feedstock powder24,25, and (3) simple mixing of additive and feedstock powder26. The research on AM from the modified feedstock material has been focused on investigating the microstructure and improving mechanical properties18,19,20. Studying the corrosion behavior of additively manufactured stainless steel from the modified feedstock has attracted only limited attention10,17,27. None of the studies reported improved corrosion performance of SLM-316L produced using modified feedstock. Additives used in feedstock modification, mostly ceramics, are envisaged to introduce electrochemical heterogeneities and therefore deteriorate the corrosion performance. On the contrary, an appropriate choice of additives that can refine the microstructure and/or suppress the formation of deleterious inclusions can be hypothesized to enhance corrosion resistance. This signifies a research gap in developing corrosion-resistant SLM-316L components.

Corrosion in stainless steel is attributed to the presence of inclusions such as manganese sulfide (MnS)3. The size and chemical composition of inclusions have been reported to control the corrosion of stainless steel28,29. Moreover, high nitrogen-containing stainless steels amassed researchers’ attention for decades for their high corrosion resistance due to nitrogen30,31,32. Therefore, stainless steels with the absence of inclusions, homogenous microstructure, and high nitrogen content are anticipated to provide high corrosion resistance. The authors hypothesize that SLM can produce highly corrosion-resistant stainless steels if the feedstock is modified using additives that suppress the formation of deleterious inclusions, homogenize microstructure, and introduce nitrogen. Therefore, chromium nitride (CrN) was selected as an additive in this research. The melting point of chromium nitride (1770 °C) is less than the SLM processing temperature, and therefore CrN should melt and be part of the matrix.

A comparison among the corrosion performance and microstructure of wrought 316L (W-316L), SLM coupons produced using commercial 316L feedstock (SLM-316L), and modified feedstock (SLM-316L/CrN) has been presented herein. The commercial-316L powders were modified by ball milling of commercial feedstock with CrN. Test coupons produced from the modified feedstock (SLM-316L/CrN) showed high corrosion performance, as evident from high pitting and repassivation potentials. Furthermore, the post-corrosion investigation revealed no metastable and stable pits in SLM-316L/CrN. Thus, this work shows a method of improving corrosion performance of additively manufactured alloys and merit in modifying the feedstock for improving properties and exploring more additives that can improve the lifespan of the AM components.

Results

Microstructural characterization

Scanning electron microscopy (SEM) of the modified feedstock (ball-milled 316LSS + 1 wt. % CrN) showed no significant change in the powder particle size and shape due to the milling process (Supplementary Fig. 1). The energy-dispersive X-ray spectroscopy (EDS) area maps showed that the CrN additive was distributed on the surface of the modified feedstock powder particles, Supplementary Fig. 1b. The density of Wrought-316L, SLM-316L, and SLM-316L/CrN using Archimedes principle were similar and calculated as 7.9 ± 0.01, 7.9 ± 0.01, and 7.89 ± 0.01 g/cm3, respectively. The X-ray diffraction analysis of SLM-316L and SLM-316L/CrN revealed the presence of only austenite phase, whereas W-316L showed austenite and along with a peak corresponding to either martensite or ferrite-phases, Supplementary Fig. 2. Literature shows possibility of martensite12,33,34 or alpha-ferrite18,35 or delta-ferrite36,37,38 in 316L stainless steel.

The backscattered electron images of SLM-316L and SLM-316L/CrN are depicted in Fig. 1. The melt pools generated from laser scanning of the individual specimen are shown in Fig. 1a, b The melt pool’s width and depth are influenced by the SLM processing parameters, selected additive and its quantity to the feedstock19,39. On the contrary, SLM-316L and SLM-316L/CrN have similar melt pool widths and depths, which indicates that the CrN additive did not influence the melt pool dimensions.

Fig. 1: SEM-BSE images at different magnifications showing the microstructure of etched SLM-316L and SLM-316L/CrN.
figure 1

a, b Cross-sectional view; c, d CCS – Columnar cellular structure and ECS – Equiaxed cellular structure; e–h ECS and CCS revealing the cells, cell boundaries, sub-cells, and sub-cell boundaries; i, j Histogram representing the area distribution of sub-cells.

Full size image

The intragranular dislocation network, which also consists of elemental cellular segregation, in short, called the cellular structure, was observed in both SLM-316L and SLM-316L/CrN. The shape of the cellular structure can be viewed as equiaxed cellular structure (ECS) and columnar cellular structure (CCS), as depicted in Fig. 1c, d. The cellular structure comprises cells (also called sub-grains) and cell boundaries (also called cell-wall/sub-grain boundaries). The near polygon-shaped cells, shown in Fig. 1e, f, are referred to as ECS39, whereas elongated cells as the CCS, as depicted in Fig. 1g, h. The width of the cell boundaries, in both ECS and CCS, were significantly higher in SLM-316L (Fig. 1e–h). A sub-cellular structure was identified inside each cell of ECS and CCS. Figure 1e–h depicts the sub-cells within each cell of ECS and CCS. The sub-cellular structure is not reported in the literature, which could be because this feature does not appear at high accelerating voltage (Supplementary Fig. 3), and the use of high accelerating voltage (20–30 kV) is a common practice. Figure 1i, j indicates the histogram representing the sub-cells area in SLM-316L and SLM-316L/CrN. The maximum number of sub-cells has an area between 0.05 and 0.15 µm2 in SLM-316L, whereas, in SLM-316L/CrN the maximum number of cells has <0.005 µm2 area. The refined sub-cell area of SLM-316L/CrN was caused by CrN addition and can be attributed to the role of CrN in providing heterogeneous nucleation sites or changing the thermal history during solidification of the melt pool40,41. Understanding the precise mechanism of sub-cell refinement due to CrN addition warrants dedicated future work.

Bright-field (BF) scanning transmission electron microscopy (STEM) image presented in Fig. 2 reveals the region of CCS of SLM-316L containing the columnar cells, cell boundaries, and sub-cellular structure. The cells were surrounded by dense entangled dislocations called cellular boundaries, whereas sub-cells were fenced by thin dislocations (interior of the cells) called sub-cellular boundaries, which can be correlated with SEM images, Fig. 1. The dislocations at the cellular boundaries and the interior of the cells have been widely reported in literature42. A series of spherical nanosized oxide-inclusions was observed at the cell boundaries and at the junction of sub-cell boundaries Fig. 2b. In addition, a tiger stripe-like pattern (termed as B-stripes) was observed in some of the cell interiors, as shown in Fig. 2c. Interestingly, no sub-cell or dislocations were noticed within B-stripe regions, whereas high dislocation densities were seen adjacent to the B-stripe pattern.

Fig. 2: STEM micrographs of SLM-316L.
figure 2

a BF image depicting the columnar cellular structure with cells and cell boundaries. b HAADF image showing the nanosized oxide-inclusions and sub-cells. c HAADF image of the B-stripe pattern.

Full size image

Figure 3 represents the STEM micrographs of CCS of SLM-316L/CrN where the columnar cells and cell boundaries are observed. A refined sub-cell size/area of SLM-316L/CrN than SLM-316L was noticed, which compliments the histograms from SEM analysis, Fig. 1e–h. Figure 3b reveals the oxide-inclusions present at the columnar cellular boundaries and the sub-cellular region.

Fig. 3: STEM micrographs of SLM-316L/CrN.
figure 3

a BF image revealing the columnar cellular structure with cells and cell boundaries. b HAADF image showing the nanosized oxide-inclusions and sub-cells.

Full size image

The STEM-EDS of SLM-316L and SLM-316L/CrN are presented in Fig. 4a, b. In both the specimens, the columnar cellular boundaries showed segregation of Cr, Ni, Mo, and S and depletion of Fe. The presence of Cr, Ni, Mo, Mn, Fe, and Si have been reported in the cellular structure of SLM-316L11,43. However, the EDS map of sulfur has not been reported in the literature. The chemical composition of nanosized oxides-inclusions in SLM-316L (30 ± 15 nm) comprises of Si, Mn, O, and S, Fig. 4a, whereas the oxide-inclusions in SLM-316L/CrN (53 ± 14 nm size) have Si, Mn, and O (Fig. 4b). Some oxide particles also contained Cr, Fig. 4b. It is worth noting that the oxide-inclusions were enriched with sulfur in SLM-316L but not in SLM-316L/CrN. The STEM analysis could not identify any nitrogen segregation in both specimens. However, the LECO ONH836 nitrogen analyzer detected 0.0820 and 0.165 wt.% nitrogen in SLM-316L and SLM-316L/CrN, respectively.

Fig. 4: STEM-EDS micrographs (HAADF).
figure 4

a SLM-316, b SLM-316L/CrN.

Full size image

Corrosion resistance

The representative cyclic potentiodynamic polarization (CPP) curves of W-316L, SLM-316L, and SLM-316L/CrN in 3.5 wt% NaCl are presented in Fig. 5a–c. The breakdown potential (Eb), repassivation potential (Erep), and maximum current density (imax) were determined from the CPP curves and presented in Fig. 5d. The forward scan of W-316L depicted in Fig. 5a shows metastable pitting initiated at ~150 mVSCE and gradually increased with potential, eventually reaching a breakdown potential (Eb) of 546 ± 20 mVSCE. The reverse scan was commenced after reaching the set current density limit of 100 µA/cm2. During the reverse scan, maximum current density (imax) of 1510 ± 62 µA/cm2 and the repassivation tendency as measured by repassivation potential (Erep) as 129 ± 13 mVSCE were obtained for W-316L. Similarly, the forward scan of SLM-316L represents the metastable pitting initiation at ~250 mVSCE and Eb at 795 ± 63 mVSCE, Fig. 5b. During the reverse scan, imax of 2276 ± 97 µA/cm2 and Erep of −113 ± 30 mVSCE were determined in SLM-316L. The forward scan of SLM-316L/CrN CPP curve, Fig. 5c, revealed no metastable pitting and higher breakdown potential (Eb = 1018 ± 38 mVSCE). The reverse scan for the SLM-316L/CrN showed significantly lower imax (104 ± 20 µA/cm2) and higher Erep (994 ± 69 mVSCE).

Fig. 5: Representative cyclic potentiodynamic polarization (CPP) curves.
figure 5

a W-316L, b SLM-316L, and c SLM-316L/CrN tested in 3.5 wt% NaCl. The inserted graphs depict the metastable pitting of the individual specimen.

Full size image

SLM-316L/CrN achieved significantly higher Eb, i.e., ~220 mVSCE (20%) higher than SLM-316L and ~470 mVSCE (45%) higher than W-316L. The imax of SLM-316L/CrN is minimal compared to SLM-316L and W-316L. The repassivation behavior measured by Erep of SLM-316L/CrN drastically improved and faster than SLM-316L and W-316L. Unlike wrought and SLM-316L, no metastable pits were identified in SLM-316L/CrN. The metastable pits developed in W-316L and SLM-316L are attributed to cause the pitting corrosion, which is evident from the lower breakdown and the repassivation potentials. The addition of CrN resulted in superior corrosion resistance noted by the absence of metastable pits, higher breakdown, and repassivation potentials, and lower current density during repassivation.

Post-corrosion surface morphology

Figure 6a–d illustrates the surface morphology after potentiodynamic polarization (PDP) of SLM-316L and SLM-316L/CrN. Figure 6a, b depicts the pit, after PDP of SLM-316L, having corroded cells and intact cell boundaries5,11. Unlike SLM-316L, SLM-316L/CrN exhibited different corroded cellular structure, Fig. 6c, d. Firstly, no pit was identified for the specimen. Secondly, the cell boundaries were corroded, and cells remained intact, Fig. 6d. This intracellular corrosion behavior could be crevice corrosion as it was only observed at the periphery of the exposed area to the electrolyte.

Fig. 6: Post-corrosion characterization after corrosion testing in 3.5 wt% NaCl.
figure 6

BSE micrographs of (a, b) SLM-316L and (c, d) SLM-316L/CrN after potentiodynamic polarization. (e) Potentiostatic polarization curves of SLM-316L and SLM-316L/CrN conducted at applied potential (Eapplied) of 700 mVSCE and 900 mVSCE, respectively. f BSE micrographs of metastable pits after potentiostatic polarization of SLM-316L.

Full size image

The potentiostatic polarization (PSP) tests were conducted for 90 min at ~100 mVSCE below Eb, i.e., applied potential (Eapplied) of 700 and 900 mVSCE for SLM-316L and SLM-316L/CrN, respectively. The current density was plotted against time, as depicted in Fig. 6e. The PSP plot of SLM-316L exhibited metastable pitting with peak current densities between ~ 60–700 µA/cm2. Stable pitting occurred after metastable pits, and an example of current transients is displayed in Fig. 6e, where eight current transients are visible. SEM characterization following PSP reveals six pits with 15–30 µm diameter, resulting in low peak current densities (60–600 µA/cm2). The two higher intensity peaks of (600–700 µA/cm2) are attributed to large pits (~200 µm diameter) shown at the right end images (orange arrows) of Fig. 6f. The PSP plot of SLM-316L/CrN did not show any metastable pitting throughout the testing period, Fig. 6e, and the SEM examination of the specimen following PSP did not reveal any noticeable corrosion. The metastable pit frequency, λ (cm−2.s−1), can be determined by normalizing the number of metastable pitting events to the product of surface area and exposure time6,44. The λ value of SLM-316L and SLM-316L/CrN are 0.063 and 0 cm−2.s−1, respectively

Passive film characterization

The passive films of SLM-316L and SLM-316L/CrN were characterized after 24 h of immersion in 3.5 wt.% NaCl using X-ray photoelectron spectroscopy (XPS) and time of flight secondary ion mass spectrometry (ToF-SIMS). The XPS peak fitting parameters, which include the binding energies and full width at half maxima (FWHM) of respective elements, are presented in Table 1.

Table 1 XPS parameters of SLM-316L and SLM-316L/CrN.
Full size table

After subtracting the Shirley baseline, high-resolution detailed XPS spectra of Fe2p3/2, Cr2p3/2, Ni2p3/2, and Mo(3d3/2, 3d5/2) were deconvoluted, Fig. 7. The survey scans of SLM-316L and SLM-316L/CrN are presented in Fig. 7a. The O1s spectra depicted in Fig. 7b is a combination of two peaks; peak-A contributions are from FeOOH, Cr(OH)3, MoO3, and H2O, whereas peak-B from Cr2O3, CrO3, Fe3O4, NiO, and MoO2. Figure 7c represents the Fe2p3/2 spectra as a combination of four constituent peaks, i.e., Fe0, FeOOH, Fe2O3, and Fe3O4. The Cr2p3/2 spectra presented in Fig. 7d comprises four peaks, i.e., Cr0, Cr(OH)3, CrO3, and Cr2O3. Researchers reported that the outer layer of the passive film mainly consists of Cr(OH)3, while the inner layer will be of Cr2O345. However, the co-existence of CrO3 and Cr2O3, due to similar standard free enthalpies, was evidenced by Olsson et al.46 and Clayton et al.47. Figure 7e presents the Ni2p3/2 spectra deconvoluted into two peaks, i.e., Ni0 and NiO. The Mo(3d3/2, 3d5/2) spectra depicted in Fig. 7f consists of six constitute peaks i.e., Mo0 (3d3/2, 3d5/2), Mo+4 (3d3/2, 3d5/2) and Mo+6 (3d3/2, 3d5/2). The Mo+4 and Mo+6 were found to be the primary oxidation states of Mo. The existence of Mo+6 as MoO42-, inhibits Cl- absorption on the sample surface and can promote the formation of Cr2O3 and CrO348,49.

Fig. 7: Passivefilm characterization using XPS.
figure 7

The high-resolution XPS spectra of (a) Survey scan (b) O1s (c) Fe2p3/2 (d) Cr2p3/2 (e) Ni2p3/2 and (f) Mo(3d3/2,3d5/2) of the passive film formed on SLM-316L and SLM-316L/CrN after immersion in 3.5 wt% NaCl at room temperature.

Full size image

The oxide/hydroxide ratios were calculated using the intensity (%) given in Table 1. The ratio of the oxides and hydroxides of Cr and Fe (given as Crox+hy/Feox+hy) for SLM-316L and SLM-316L/CrN is 2.66 and 2.85, respectively. Cr-oxides are Cr2O3 and CrO3, Cr-hydroxide is Cr(OH)3, Fe-oxides are Fe2O3 and Fe3O4, and Fe-hydroxide is FeOOH. Also, the total oxide (Cr2O3, CrO3, Fe2O3, Fe3O4, NiO) to hydroxides (Cr(OH)3 and FeOOH) ratio given as Totalox/Totalhy of SLM-316L and SLM-316L/CrN are 3.58 and 3.03, respectively. These ratios indicate the higher participation of chromium oxide in the passive film.

Figure 8 presents the ToF-SIMS negative ion depth profiles of SLM-316L and SLM-316L/CrN. The intensities of secondary ion characteristics of the passive film (18O−, OH−, FeO2−, CrO2−, NiO− and MoO2− shown in Fig. 8a, c) and substrate (Fe2−, Cr− and Ni− shown in Fig. 8b, d are plotted in logarithmic scale for individual specimens. Similar depth profiles of oxide and metallic ions are observed in both specimens. The interface, represented by bold dotted line, was determined at a particular sputter time where a significant change in slope of metallic peaks was observed50,51. A gray dotted line (~25 s) was drawn where the hydroxide and oxide zones can be bifurcated. The hydroxide and oxide zones can be represented in zone-1 and zone-2, respectively. The oxygen characteristic of the oxide scale is explained by 18O− as 16O− saturates the detector. High-intensity peaks of OH−, FeO2−, MoO2− were found in zone-1, whereas NiO− are found in zone-2, Fig. 8a, c. The high-intensity peaks of OH− and FeO2− indicate a high probability of Fe-hydroxide (FeOOH) formation in zone-1, i.e., the outermost region of passive layer48,52. The high-intensity peak of MoO2− profile found in zone-1 may indicate that oxides of Mo are enhancing the stability of the passive film49,50,53. The CrO2− profile intensity is uniform almost throughout the passive layer, Fig. 8a, c, which agrees with XPS results that chromium has higher participation in the passive film, Fig. 8d. The enrichment of Ni− metallic ions at the interface between passive layer and metallic substrate was observed51,53, Fig. 8b, d.

Fig. 8: Passivefilm characterization using ToF-SIMS.
figure 8

ToF-SIMS negative ion depth profiles of (a, b) SLM-316L and (c, d) SLM-316L/CrN after immersion in 3.5 wt% NaCl for 24 h at room temperature.

Full size image

As evidenced from XPS and ToF-SIMS, both the SLM-316L and SLM-316L/CrN exhibited similar passive film compositions. Therefore, the possible role of change in passive film composition in improving corrosion resistance can be overruled.

Discussion

The pitting corrosion resistance (Fig. 5) was in the order of SLM-316L/CrN > SLM-316L > W-316L. The higher corrosion resistance of SLM-316L in comparison to W-316L can be attributed to the absence of MnS. It is well known that the manganese sulfide (MnS) inclusions are the common impurities found in wrought and are the reason for pitting corrosion4. The nucleation frequency and metastable pitting have been reported to depend on the size and distribution of MnS particles54,55. A critical MnS particle size of ~0.7 µm is required for stable pit formation28,29. The rapid solidification rate during the SLM does not provide enough time for the elements to diffuse and form MnS inclusions6. Therefore, the absence of MnS inclusions in SLM-316L, confirmed by the SEM and STEM analysis, is attributed to higher Eb than wrought-316L, which is in agreement with the literature6. However, other sulfur-rich inclusions can act like MnS inclusions causing corrosion. Ke et al.29 had reported three types of inclusions in austenitic stainless steel: MnS, the existence of multielement (Al, Mn, Cr) oxides, and a mixture of sulfides and oxides inclusions. It was stated that the multielement oxides and combination of sulfide and oxide inclusion could initiate corrosion29,54. Although MnS inclusions were absent in SLM-316L, the next possible sulfur-rich inclusions (Si-Mn-S-O), as evident from STEM (Fig. 4a), is attributed to act as potential cathodic site and initiate pitting corrosion. The chemical inhomogeneity in the cellular structure favored transforming the metastable pitting into a stable pit. The cells having depletion of solute were preferentially targeted, causing superior corrosion within cells than cell boundaries, Fig. 6a, b. The absence of sulfur-rich inclusions in SLM-316L/CrN was attributed to the elimination of metastable pitting followed by stable pitting corrosion, Fig. 6e. This eventually favored higher Eb, and the inability to cause pitting corrosion might have supported the faster repassivation ability.

The improvement in corrosion performance of SLM-316L/CrN can also be attributed to increased nitrogen%. The nitrogen% of SLM-316L and SLM-316L/CrN is determined to be 0.0820 and 0.165 wt.%, respectively. Many researchers reported enhanced pitting resistance, repassivation ability, and reduced metastable pitting of high nitrogen induced 316 L in chloride solution32, sodium sulfate solution56, and thiosulfate-chloride solution31. Several mechanisms were suggested to explain the phenomena: 1. Nitrogen in solid solution promotes NH4+ formation, which can reduce oxidation inside the pit31 2. Improving the passive film stability by nitrogen enrichment and preventing the attack of Cl− ions57 3. Producing nitrate ions which can improve the pitting corrosion resistance58 4. Nitrogen alloying enriched the chromium% within the passive film32. The presence of NH4+ commonly found in the passive layer was evidenced to improve the corrosion behavior30,31,56,59. Lee et al.32 studied the effect of nitrogen% on metastable pitting susceptibility of 316LSS using PSP experiments. The authors reported the decreased metastable pitting with increased nitrogen%, and no metastable pitting was observed in specimens having nitrogen% ≥0.151 wt.%. A high acidic environment (low pH) favors faster metal dissolution and initiating metastable pits. Nitrogen in solid solution can combine with H+ forming NH3 and NH4+ thus, increasing the local pH31. Nitrogen-induced 316L having higher local pH promoted higher pitting potential and faster repassivation31,32. A similar interpretation can be obtained for SLM-316L/CrN. From XPS and SIMS analysis, no evidence of nitrogen active compound and ion was detected, respectively, which could be attributed to low nitrogen% in SLM-316L/CrN. Dai et al.31 investigated corrosion performance of 316LSS with different nitrogen content and reported that the N1s XPS high-resolution spectra of passive film formed on 316LSS with ≥0.2 N% comprised peaks of NH3 and NH4+.

The chemical composition of specimens determined using inductively coupled plasma optical emission spectrometry, is presented in Table 2. Higher percentage of Mo, Ni, and Cr in SLM-316L/CrN than SLM-316L was observed. The increased Cr content in SLM-316L/CrN can be expected from CrN addition. Besides, the CrN addition might be effective in restricting the evaporation of Ni and Mo during SLM. The elements like Mo, Ni, and Cr are known for inhibiting pitting corrosion in austenitic stainless steels by favouring a stable passive film formation51,60,61,62.

Table 2 Chemical compositions in wt.% of W-316L, SLM-316L and SLM-316L/CrN.
Full size table

In summary, the corrosion performance of SLM of 316L was significantly improved by feedstock modification using the CrN additive due to modification of microstructure and chemical composition. Similarly, several other additives, most importantly nitrogen-containing, have the potential to create a highly corrosion-resistant stainless steel alloy using AM. This work shows merit in exploring more additives and further understanding the role of CrN addition in enhancing corrosion resistance.

Methods

Feedstock modification and sample manufacturing

The gas-atomized 316L stainless steel powders were procured and termed commercial feedstock herein. The commercial feedstock has particles with maximum dparticle = 63 µm, and its chemical composition is provided in Table 2. The wrought 316L (as W-316L) sheets are purchased from McMaster. Chromium nitride (CrN) additive with dparticle = 1–3 µm was purchased from US Research Nanomaterials. The composite feedstock powders were produced by ball-milling of commercial feedstock along with 1wt.% CrN. The ball-milling was carried out using stainless steel balls and jars at 180 RPM for 5 h with a 1:1 ball-to-powder ratio (BPR) in a planetary ball mill.

The selective laser melting (SLM) was conducted using the Renishaw AM400 laser powder bed fusion system, employed to 400 W ytterbium fiber laser with 70 µm laser diameter. The test specimens of commercial and composite feedstock were printed in an argon atmosphere with an oxygen setpoint <2000 ppm, and the building parameters are given in Supplementary Table 1. The printed specimens are cylinder in shape with 15 mm diameter and 5 mm height. In this research, the SLM printed specimens using commercial and modified feedstock are termed SLM-316L and SLM-316L/CrN, respectively.

Characterization

The chemical composition of SLM specimens was determined using an Agilent inductively coupled plasma optical emission spectrometer (ICP-OES), LECO carbon/sulfur analyzer and LECO ONH836 oxygen/nitrogen/hydrogen analyzer following the AL0025 and ASTM E1019 standard procedure. The density of the specimens was measured following Archimedes’ principle. X-ray diffraction (XRD) of SLM specimens was conducted using Rigaku smartlab X-ray diffractometer with graphite monochromatized Cu-Kα radiation (λ = 0.154056 nm).

Scanning electron microscopy (SEM) using FEI Verios 460L field emission electron microscope was performed for microstructural characterization. The SLM specimens were ground to 1200-grid SiC sandpaper and later fine polished till 0.05 µm. After polishing, the microstructures were observed, followed by electro-etching with 10% oxalic acid at 15 V for 60 s (as per the ASTM A262 practice- A). The sub-cell area was measured using ImageJ software.

Scanning transmission electron microscopy (STEM) was performed using Talos F200X G2 scanning transmission electron microscope operating at 200 kV. The STEM High-angle annular dark-field (HAADF) and bright-field (BF) imaging of SLM-316L and SLM-316L/CrN TEM lamellae were conducted along with SuperX energy-dispersive x-ray spectroscopy (Super-X EDS) recorded. The STEM sample preparation was performed using the Focused Ion Beam (FIB) technique. The electropolished samples were loaded into the Quanta 3D FEG Focused Ion Beam microscope to prepare site-specific TEM lamellae from SLM-316L and SLM-316L/CrN specimens. The Quanta FIB is a dual-beam instrument combining an electron beam and ion beam for simultaneous imaging and milling. The specimens were aligned according to the feature of interest; later, ~2 µm platinum cap was deposited on the surface to protect it from ion beam damage during high-current ion beam milling. After the cross-sections were ~ 2 µm thick, they were cut and placed on TEM grids using the Omniprobe micromanipulator, enabling in-situ TEM sample preparation. The thinning down and polishing of the cross-sections were completed using low current and voltage to avoid sample amorphization and damage, and then electron transparent TEM specimens were taken for TEM characterization.

The passive film of SLM-316L and SLM-316L/CrN was developed following immersion in 3.5 wt.% (0.6 M) NaCl for 24 h. The passive film of the immersed specimens was characterized using X-ray photoelectron spectroscopy and secondary ion mass spectroscopy.

The chemical composition of the passive films formed over the SLM specimen’s surface was quantitatively investigated using SPECS X-Ray photoelectron spectroscopy (XPS) with 300 W Mg-anode (hν = 1253.6 eV) X-Ray source. The experiment was performed at a 60° take-off angle using a 2 × 2 mm2 scan area. First, survey spectra were recorded from 0-1100 eV binding energy with a 0.5 eV step size to identify all the elements present on the passive film. Later, the high-resolution regional spectra of O1s, Fe2p, Cr2p, Ni2p, and Mo3d with 0.1 eV step size were recorded to comprehend the passive film behavior and their respective oxidation states. Deconvolution of peaks from XPS spectra was performed using Casa XPS software after Shirley-type background subtraction. The binding energies were calibrated with the adventitious carbon C1s peak position at 285 eV.

Elemental depth profiling on all the SLM passive film developed specimens were carried out using time-of-flight secondary ion mass spectroscopy (ToF-SIMS) operated under 5 × 10−9 mbar pressure. A pulsed 25k eV Bi3+ primary ion source delivering 0.32pA of target current over 100 × 100 µm2 area was used for analysis. Depth profiling was performed by interlacing secondary ion analysis with sputtering using a Cs+ ion beam with 1 keV giving 7.2 nA target current over 200 × 200 µm2 area. Edge effects were avoided by centering the analysis inside the eroded crater. The profiles were recorded with negative secondary ions due to the high sensitivity to ion yield of the oxide film.

Corrosion tests

Cyclic potentiodynamic polarization (CPP) of the W-316L, SLM-316L, and SLM-316L/CrN was conducted in 3.5 wt% NaCl electrolyte at room temperature on the specimens metallographically prepared to 1200-grit SiC grinding. A three-electrode flat cell having a saturated calomel reference electrode (SCE) and a platinum mesh counter electrode (CE) was used for corrosion testing. The CPP tests were initiated at 0.20 VSCE below open circuit potential (OCP), and a scan rate of 1 mV/s was used. The forward scan was terminated, and a reverse scan was initiated when either the potential reached 1.5 VSCE or a current density reached 100 µA/cm2. Before polarization, open circuit potentials were recorded while the specimens were stabilized in test electrolyte for 1 h. The breakdown potential (Eb), repassivation potential (Erep), and maximum current density (imax) were determined from the CPP curves, which were used to compare the corrosion performance of the tested specimens.

Potentiodynamic polarization (PDP) and potentiostatic polarization (PSP) experiments were conducted on SLM-316L and SLM-316L/CrN, and the post-corrosion investigation was performed on the tested specimens. The PDP test was conducted similarly to the CPP test, except the forward scan was ceased when either the potential reached 1.5 VSCE or a current density reached 1000 µA/cm2, and no reverse scan was initiated. The PSP tests of SLM-316L and SLM-316L/CrN were conducted at applied potential (Eapplied) of 700 and 900 mVSCE, respectively, to an exposed area of 0.079 cm2, for 90 min. Later, the PDP and PSP tested specimens were taken under SEM to observe metastable and stable pit morphologies.