Additively Manufactured 316L Stainless Steel Subjected to a Duplex Peening-PVD Coating Treatment

This research studies the individual and combined effects of mechanical shot peening and the deposition of TiAlCuN coating on additively manufactured 316L stainless steel. Shot peening has been found to induce a 40% increase in surface hardness, while the combined effect of shot peening and the coating produced an approximately three-fold increase in surface hardness when compared to the as-printed coupons. Shot peening reduced the surface roughness of printed metal coupons by 50%, showing that shot peening can also serve to improve the surface finish of as-printed 316L stainless steel components. The peening process was found to induce a compressive residual stress of 589 MPa, with a maximum affected depth of approximately 200 μm. Scratch testing of the printed and coated specimens showed complete delamination failure at a normal load of 14 N, when compared to hybrid treated samples which failed at 10 N. On the other hand, from the corrosion tests, it was found that the hybrid treated samples provided the optimal results as opposed to the other variables.


Introduction
The marine transportation industry makes up part of the global economy and trade, since primary necessities such as petroleum, foods and goods are carried by means of water. Due to their working environment, parts such as propellers and shafts are exposed to the adverse effects of corrosion, erosion, wear, and detrimental effects. The failure of the fundamental components of a ship can lead to severe consequences, mainly delays in delivery of cargo, environmental damage, financial losses and most importantly fatalities of the passengers on board [1]. Such mid-journey failure would dictate a potentially long sourcing processing and docking of the vessel. The high demand for replacement of parts could be sustained by additive manufacturing (AM).
Components suitable for the marine environment necessitate a particular combination of attributes to make them suitable for use, first and foremost hardness. Additive manufactured 316L stainless steel (SS) is known to have a higher hardness than wrought 316L SS. In a study by Yusuf et al. [1], an average microhardness value of 228 HV was obtained for AM 316L SS, while 192 HV was measured for wrought 316L SS [1]. This increase in microhardness is attributed to the finer grains within the microstructure [2]. Additionally, parts intended for engine and drive components need to possess a high tensile and yield strength to be able to withstand the high loads experienced during use. AM can produce a wide range of ultimate tensile strengths ranging from 550 to 700 MPa, while values for the yield strength vary from 300 to 600 MPa, depending on the chosen parameters used for printing the material [3][4][5][6]. Similarly, corrosion resistance is also an important characteristic for marine transportation parts due to the environment they are subjected to. Few studies on the corrosion resistance of additively manufactured 316L SS are available and Table 1. Main elements of the chemical composition of the wrought 316L SS and the powder used for AM [26,27].

Tensile and Impact Testing
Tensile and impact tests were carried out to study the bulk tensile strength and toughness of the material. The tensile samples having dimensions shown in Figure 1 were printed according to the standard ASTM E8/E8M-16a: Standard for Tension Testing of Metallic Materials [28]. The sub-sized dimensions were chosen so that the samples could be printed within the printer's limited build volume. The tensile tests were carried out along the x-direction using an Instron Universal Testing System 5892 (Norwood, MA, USA) with the attachment of an Instron extensometer 2530 (USA). A strain rate of 1 mm per minute was used until the elastic region was exceeded and was then changed to 3 mm per minute after removing the extensometer, which strain rate was kept until fracture occurred. Six repeated tests were carried out.    The Charpy impact testing samples having dimensions of 55 × 10 × 10 mm, were manufactured according to the standard ASTM E23-18: Standard Test Methods for Notched Bar Impact Testing of Metallic Materials [29], as shown in Figure 2. The impact testing was performed using an Instron 450MPX-J2 (USA) motorised pendulum impact testing system and a velocity of 5.32 m/s. The reported results are the average of six measurements.

Shot Peening and Coating Deposition
Shot peening was carried out using a modified set up in an Industrial Surface Treatments Ltd. AB850 air blasting machine and S230 shots. A nozzle of 80 mm length and 7 mm diameter, nozzle to specimen distance of 100 mm, pressure of 7 bar and an Almen intensity of 0.21 mmA were used.
The coating deposition was carried out using a Teer UDP800 (Beijing, China) closed field unbalanced magnetron sputtering ion plating machine. The system was composed of 2 titanium targets, 1 aluminium and 1 copper target, with an argon and nitrogen gas inlet. Table 2 shows a summary of the coating deposition parameters.

Shot Peening and Coating Deposition
Shot peening was carried out using a modified set up in an Industrial Surface Treatments Ltd. AB850 air blasting machine and S230 shots. A nozzle of 80 mm length and 7 mm diameter, nozzle to specimen distance of 100 mm, pressure of 7 bar and an Almen intensity of 0.21 mmA were used.
The coating deposition was carried out using a Teer UDP800 (Beijing, China) closed field unbalanced magnetron sputtering ion plating machine. The system was composed of 2 titanium targets, 1 aluminium and 1 copper target, with an argon and nitrogen gas inlet. Table 2 shows a summary of the coating deposition parameters.

Material Characterisation
Scanning electron microscopy (SEM) was performed by a Carl Zeiss Merlin field emission scanning electron microscope (Oberkochen, Germany) having a Gemini II column. Electron dispersive spectroscopy (EDS) was performed by means of an Ametek EDAX Apollo X 2189 (Mahwah, NJ, USA) attachment located within the SEM. This was used to obtain the chemical composition and elemental maps of the studied surfaces. X-ray diffraction (XRD) phase analysis of all the specimens was carried out using a Rigaku Ultima IV X-ray diffractometer (Tokyo, Japan), equipped with crossbeam optics (CBO) set in Glancing Angle Incidence Asymmetric Bragg (GIAB) configuration, with a 3 • angle of incidence and a scan range between 20 • and 120 • .
Residual stress measurement at the surface and near the surface up to a depth of 200 µm was carried out using the XRD, according to the standard BS EN 15305 (2008)-Non-destructive testing-Test method for residual stress analysis by X-ray diffraction [30], using the sin 2 ψ method. The measurement for peak shifting for the calculation of the residual stresses was performed on the (311) austenite peak at a 2θ value of 90.68 • . Each surface was tilted at seven different ψ angles between 0 and 60 • . The θ-2θ scans were performed between 2θ values of 87 • and 93 • . An electrolyte consisting of 5.4% perchloric acid, 94% ethanol and 0.6% de-ionised water was used to carry out electropolishing on a Struers LectroPol-5 electrolytic polishing machine (Copenhagen, Denmark) to progressively remove layers of the material.
Surface micro-hardness tests were executed on a Mitutoyo MVK-H2 micro-hardness testing machine (Kawasaki, Japan), equipped with a Vickers pyramidal indenter and loaded with a 200 gf indentation load, having a dwell time of 10 s. For coating evaluation, nanoindentation tests were performed on a mirror-finished coated wrought 316L SS sample. A Nanomaterial NanoTest 600 machine (Wrexham, UK) was equipped with a 18580-a Berkovich 120 • diamond tip indenter. The maximum indentation load was set to 50 mN. Thirty indentations were made on the specimen, spaced at 30 µm from each other.
The surface roughness was measured using an AEP Technology NanoMap 500 LS 3D contact profilometer (California, USA), equipped with a 1 µm stylus tip. The results presented are an average of five measurements.

Scratch Testing
A UMT Bruker TribolabTM tribometer (San Jose, CA, USA) having a 60 • Rockwell type C indenter was used to execute micro-scratch testing. A ramped load starting from 0.5 N up to 40 N was used, with a force of 1.33 N/s and a scanning velocity of 0.339 mm/s. Five scratches of 10 mm each were done on the sample. The scratch morphology was then studied under the optical microscope and SEM to identify the positions and loads at which failure took place.

Corrosion Studies
Cyclic polarisation tests were performed on the cylindrical samples using a threeelectrode setup to study the corrosion response of the material. The three-electrode setup was connected to a Gamry Interface 1000 TM potentiostat (Warminster, PA, USA), having the sample as the working electrode, a platinum coated titanium rod as the counter electrode and a saturated calomel electrode (SCE) as the reference electrode. Then, 300 mL of testing solution was prepared according to ASTM D 1141-98: Standard Practice for the Preparation of Substitute Ocean Water [31], where 1 cm 2 of surface area was exposed to the electrolyte. An initial OCP test of 2 h was performed, followed by cyclic polarisation sweeps at a rate of 0.167 mV/s, which was reversed at an apex current density of 0.5 mA/cm 2 . Three repeats were performed. Table 3 shows the samples which were tested in this study, together with respective abbreviations used throughout this article.

Error Calculation
The data presented in this work are the sample mean (x) value obtained from the measured values, with a sample size (n), specified in each section. The quantitative data presented in graphical formats have been included with error bars, while that presented in numerical formats has been included with a ±value. This was done to ensure the correct statistical interpretation of the data. Since most of the sample sizes were smaller than 30, the error ranges were calculated using the t-distribution [32].  Table 4. The measured Young's Modulus, the yield strength, UTS and the elongation all fall within the ranges found in the available literature for additive manufactured 316L SS. While the UTS of the AM samples is similar to that found in literature for wrought 316L SS, some discrepancies are evident for Young's Modulus, yield strength and elongation values. A 19% decrease in Young's Modulus is evident between AM SLM and conventionally manufactured 316L SS. Similar values were obtained by Merkt [41] with a Young's Modulus of 140 GPa on AM 316L SS. This could be attributed to the parameters of the 3D printing process, specifically the build-up direction and laser power used, which provide different crystallographic orientation of the grains [33,42]. Niendorf et al. [42] report that the grains have a preferential crystallographic orientation according to the laser power utilised such that the grains were oriented in the (011) direction when a laser power of 400 W was used, while the grains were oriented in the (001) direction with a laser power of 1000 W [43]. In austenitic SS, the preferential orientation is that of (001), giving a decreased Young's Modulus [42]. In addition, the porosity present in additive manufactured materials also results in a decrease in the Young's Modulus. An increase of 60% in the yield strength, 4% in UTS and 10% in elongation can be noted for AM SLM over wrought. The significant improvement in YS can be attributed to refined microstructure obtained during the high cooling rates of the SLM process. Additionally, the high yield strength and elongation are attributed to high dislocation densities and twinning formations during the SLM process, respectively. Tensile stresses in SLM materials also lead to a high yield strength, with the high dislocation densities formed during deformation [36,44]. The small increase of UTS indicates that during testing, SLM materials do not exhibit the same amount of work hardening as the wrought.

Impact Tests
A total impact energy, resulting in impact toughness, of 75 ± 2 J was obtained (Table 5) after fracturing the sample ( Figure 2). This value falls within the range found in literature for AM SLM 316L SS. This value is slightly lower than that of wrought 316L stainless steel. Table 5. Impact properties.

Maximum Load (kN) Total Energy (J)
AM SLM (Measured) 15 ± 0.10 75 ± 2 AM SLM (Literature) [45][46][47][48] / 60-100 Wrought (Literature) [45][46][47][48] 120-180 A micrographic analysis of the fractured surface was carried out, as shown in Figure 3a-c. The material which has experienced compressive loading during impact testing shows a distinctively flattened morphology, whereby the material texture both that formed by plastic deformation and features characteristic of AM were compressed against each other, as shown in Figure 3a. In fact, cross-sectional evaluation has revealed some residual porosity of less than 0.2% in volume. Ductile deformation is evidenced by the cup-andcone structure shown in Figure 3c and pores (Figure 3b), cracks and dimples. The pores, a characteristic of ductile fracture, are formed due to insufficient bonding of melt pools which are next to each other, during the solidification processing [49,50].  Figure 4 shows the microstructure of the AP 316L SS composed of an austenitic matrix. At higher magnifications, shown in Figure 5, columnar and cellular dendritic structures were observed. Such structures are formed following molten metal solidification. Additionally, Figure 6a-d respectively show optical micrographs of the surface of AP, PSP, PC and PSPC specimens. These micrographs show the 100% coverage produced by the shot peening. Figure 6b,d show the individual SP dimple characteristics. SP generated a less rough surface than the as-printed, as will be observed later in Section 3.3.   Figure 6b,d show the individual SP dimple characteristics. SP generated a less rough surface than the as-printed, as will be observed later in Section 3.3.        Figure 7 shows a comparison between the AP, shot peened, coated and hybrid treated coupons. The peening treatment resulted in a 50% reduction for Ra and 80% for Rz and can be attributed to the fact that as the shots impinge the surface, the rough crests resulting from the printing process are compressed, the protrusions on the surface are deformed radially, forming individual dimples which flatten the surface and thus, reduces the roughness. The surface roughness reduction by SP was also reported by Sugavaneswaran et al. [15] where a 50% deduction in the average surface roughness was discovered after SP AM 316L SS, using S390 shots with a 1 mm diameter, for 15 min, and with a 200% coverage.  Figure 7 shows a comparison between the AP, shot peened, coated and hybrid treated coupons. The peening treatment resulted in a 50% reduction for Ra and 80% for Rz and can be attributed to the fact that as the shots impinge the surface, the rough crests resulting from the printing process are compressed, the protrusions on the surface are deformed radially, forming individual dimples which flatten the surface and thus, reduces the roughness. The surface roughness reduction by SP was also reported by Sugavaneswaran et al. [15] where a 50% deduction in the average surface roughness was discovered after SP AM 316L SS, using S390 shots with a 1 mm diameter, for 15 min, and with a 200% coverage.   Figure 8 portrays the XRD diffractographs for the AP, shot peened, coated and hybrid treated coupons. The main differences identified between the AP and shot peened diffractographs were: (i) change in relative intensity at the (111) and (200) peaks, (ii) a poorer definition of the (110) ferrite peak, (iii) broadening of XRD peaks in the shot peened sample and (iv) a slight peak shift for the (200) peak. The broadening and shifting of the XRD peaks are attributed to the macro and micro residual stresses induced by SP [51,52]. Similar differences between the printed and coated, and the hybrid samples were obtained, including: (i) change in relative intensity in the (111)

XRD Stress Evaluation
A surface residual stress of 61 ± 4 MPa, −589 ± 6 MPa and −693 ± 8 MPa was obtained for the as-printed, printed and shot peened and printed, shot peened and coated When analysing the diffractographs of the coated and hybrid treated, the peaks obtained agree with those obtained by Man et al. [53] when studying TiAlN thin films. The two peaks of (111) and (200) for TiN are in the same position as austenite. The XRD technique did not detect any Al and Cu crystalline compounds with the two elements having diffused to form a solid solution. The XRD patterns of the coated surfaces are different from those of the substrate, confirming that there are distinct phases of the coating, even though some of the austenite peaks were still present.

XRD Stress Evaluation
A surface residual stress of 61 ± 4 MPa, −589 ± 6 MPa and −693 ± 8 MPa was obtained for the as-printed, printed and shot peened and printed, shot peened and coated respectively. Figure 9 portrays the residual stresses developed along the depth for the as-printed (AP) and shot peened (PSP) samples.  The AP samples exhibited tensile stresses of around 61 MPa, both at the surface and the sub-surface. Tensile stress values have been associated with several mechanisms including: (i) temperature gradient, (ii) re-melting and solidification of layers and (iii) inhomogeneous lattice spacing [54]. On the other hand, the SP treatment induced a maximum compressive residual stress of around 589 MPa. Figure 10 shows that the depth of the shot peened layer is around 250 μm. This is in line with other work on AM 316L SS carried out by Gundgire et al. [16], where the affected depth was in the range of 225 to 275 μm.
Compressive residual stresses for the hybrid treated specimens were similar to those obtained on the shot peened coupon. A surface stress of −693 ± 7 MPa was achieved which reached −561 ± 5 MPa at an affected depth of 54 μm. This outcome shows that the coating deposition on the peened AM specimen did not remove the beneficial compressive stresses induced by the peening process.  The AP samples exhibited tensile stresses of around 61 MPa, both at the surface and the sub-surface. Tensile stress values have been associated with several mechanisms including: (i) temperature gradient, (ii) re-melting and solidification of layers and (iii) inhomogeneous lattice spacing [54]. On the other hand, the SP treatment induced a maximum compressive residual stress of around 589 MPa. Figure 10 shows that the depth of the shot peened layer is around 250 µm. This is in line with other work on AM 316L SS carried out by Gundgire et al. [16], where the affected depth was in the range of 225 to 275 µm. The AP samples exhibited tensile stresses of around 61 MPa, both at the surface and the sub-surface. Tensile stress values have been associated with several mechanisms including: (i) temperature gradient, (ii) re-melting and solidification of layers and (iii) inhomogeneous lattice spacing [54]. On the other hand, the SP treatment induced a maximum compressive residual stress of around 589 MPa. Figure 10 shows that the depth of the shot peened layer is around 250 μm. This is in line with other work on AM 316L SS carried out by Gundgire et al. [16], where the affected depth was in the range of 225 to 275 μm.
Compressive residual stresses for the hybrid treated specimens were similar to those obtained on the shot peened coupon. A surface stress of −693 ± 7 MPa was achieved which reached −561 ± 5 MPa at an affected depth of 54 μm. This outcome shows that the coating deposition on the peened AM specimen did not remove the beneficial compressive stresses induced by the peening process.  Compressive residual stresses for the hybrid treated specimens were similar to those obtained on the shot peened coupon. A surface stress of −693 ± 7 MPa was achieved which reached −561 ± 5 MPa at an affected depth of 54 µm. This outcome shows that the coating deposition on the peened AM specimen did not remove the beneficial compressive stresses induced by the peening process. Table 6 shows that shot peening treatment improves the hardness of the as-printed material by 40%, from 238 HV to 334 HV. This is in line with the study by Gundgire et al. [16], where a hardness of 340 to 360 HV was achieved after SP AM 316L SS. This increase is attributed to plastic deformation taking place during SP. The intrinsic hardness of the coating, which was measured by nanohardness, shows that it further improves the characteristics of the surface of the material. Then, the combination of the shot peening and the coating treatment provides an even superior compound value of hardness, giving 2.9 times increase in hardness over the as-printed samples.  Figure 10 showcases the microhardness depth profile of the shot peened and hybrid sample. It can be noted that the affected depth is also around 250 µm, which is in line with the affected depth obtained in the residual stress measurement (Section 3.5). The affected depth is comparable with that of 189 µm and 225-275 µm obtained by Maamoun et al. [55] and Gundgire et al. [16], respectively.

Adhesion Tests
The coating and the hybrid treatment showed a similar behaviour of coating characteristics following scratch testing. Figure 11 shows that L C2 and L C3 were detected along the wear track of coated, while for the hybrid treated L C3 only was identified. L C1 could not be identified for both variables. The L C2 characteristic was made from initial delamination of the coating, while the L C3 characteristic was made from interfacial shell-shaped spallation. The earliest sign of adhesive failure and substrate exposure was noted on the hybrid at a distance of 2.6 ± 0.34 mm, and as shown in Figure 11, this corresponds to a scratching load of 10 ± 1 N. This is in contrast with the results obtained in a study by Tillmann et al. [54], in which a CrAlN coating was deposited on AM 316L SS. Both L C2 and L C3 were obtained at a force of 7 ± 1.7 N and 38.4 ± 3.5 N, respectively. This shows that the CrAlN coating provided better adhesion to the substrate as it failed at higher loads.
Further mixture of cohesive and adhesive failure was identified throughout the wear track of coated and the hybrid treated, showing more delamination and interfacial shell spallation, as shown in Figures 12c and 13c. These were formed as the scratch load increased. Such characteristics were elevated since the soft material was not able to properly support the coating from cracking and forming such defects. As observed in Figure 13c, the perforation of the scar did not result in total coating delamination, even as the load was increased. These were replaced by interval delamination because of residual stress relaxation during coating spallation. Materials 2023, 16, x FOR PEER REVIEW 15 of 25 Figure 11. Critical loads achieved at ramped loads for the hybrid, printed and coated and wrought and coated.    Additionally, the scratch testing on the wrought and coated was performed to serve as a control to the coated and the hybrid. During this testing, all the three adhesion characteristics were identified. At low loads, LC1 was shown, which was characterised by forward chevron cracks, longitudinal to the scratch track, showing cohesive failure (Figure 14b). Adhesive failure was then identified at a load of 12 ± 0.2 N at which load the coating delaminated along the scratch track (Figure 14c). At further higher loads, LC3 (Figure 14d) took place at a load of 17 ± 0.4 N. This characteristic consisted of full interfacial shell spallation, with full delamination of the coating taking place longitudinally along the scratch track. Chevron cracks and localised chipping were found along the track.
When comparing the values in Figure 11, the PC performed better than the PSPC evidenced by the first failure mode detected at a higher load than that measured on the hybrid equivalent. The similar nature of the coated chemical makeup at the surface suggests that the difference in performance can be attributed both to the test mechanics, where the tip interaction changes with the roughness of the sample being measured and the improved load support provided by the harder and stiffer coating. The results of the PC are also superior to the wrought and coated, since the first failure on the wrought and coated was seen at an earlier load. The wrought and coated tests were performed to analyse all the three characteristics synonymous with scratch testing. Additionally, the scratch testing on the wrought and coated was performed to serve as a control to the coated and the hybrid. During this testing, all the three adhesion characteristics were identified. At low loads, L C1 was shown, which was characterised by forward chevron cracks, longitudinal to the scratch track, showing cohesive failure (Figure 14b). Adhesive failure was then identified at a load of 12 ± 0.2 N at which load the coating delaminated along the scratch track (Figure 14c). At further higher loads, L C3 (Figure 14d) took place at a load of 17 ± 0.4 N. This characteristic consisted of full interfacial shell spallation, with full delamination of the coating taking place longitudinally along the scratch track. Chevron cracks and localised chipping were found along the track.

OCP Curves
The OCP curves for the wrought, as-printed, polished, shot peened and hybrid treated followed a similar behaviour, until they stabilised for the rest of the curve's When comparing the values in Figure 11, the PC performed better than the PSPC evidenced by the first failure mode detected at a higher load than that measured on the hybrid equivalent. The similar nature of the coated chemical makeup at the surface suggests that the difference in performance can be attributed both to the test mechanics, where the tip interaction changes with the roughness of the sample being measured and the improved load support provided by the harder and stiffer coating. The results of the PC are also superior to the wrought and coated, since the first failure on the wrought and coated was seen at an earlier load. The wrought and coated tests were performed to analyse all the three characteristics synonymous with scratch testing.

Corrosion Tests OCP Curves
The OCP curves for the wrought, as-printed, polished, shot peened and hybrid treated followed a similar behaviour, until they stabilised for the rest of the curve's duration. This shows that the setup had stabilised and was ready to carry out the cyclic polarisation test. However, the same cannot be said for the coated curve. This is due to the fluctuations in the curve taking place. The reason for these fluctuations could be due to metastable pits growing but their growth is stopped abruptly and repassivated. This repeatable behaviour obtained from a set of repeats for each set, provided an early indication of how the sample was going to perform in the cyclic polarisation test, providing poor corrosion results, performing the worst with respect to the six samples tested.
Cyclic Polarisation Curves Figure 15 shows the representative curves for the cyclic polarisation tests, while Table 7 provides a summary of important numerical values extracted from the plot and repeats. The most noble E corr was obtained by the AP, whilst the most negative was achieved by the PC sample. The rest of the samples had very similar E corr ranging from −210 to −180 mV. The more noble an E corr is, the lower the corrosion susceptibility [56] and the higher the stability of the passive film [7].
The highest E break was obtained for PP at 776 ± 127 mV, whilst the least was achieved by the PC sample at 241 ± 80 mV. The range for the other samples is between 330 and 690 mV. The E break , also known as the pitting potential, is the lowest potential at which the material will succumb to pitting corrosion. Above the E break , new pits will form [57]. Therefore, the higher the E break , the more resistance to pitting and the further improved stability of the oxide film [56,58].
A fluctuating current density in the passive region of the anodic scans shows the formation of metastable pits [7]. This was showed by the wrought, as-printed, shot peened and coated samples. Their growth is stopped rapidly and repassivated [59]. This repassivation takes place with the aid of the salt films formed by the electrolyte. They suppress the transfer of cations and more growth of the pits. From the curves in Figure 15b, it can be noted that the polished sample showed the lowest metastable pits formation, while the as-printed and the coated showed the most metastable pits formation, with the most fluctuations below E break . From this analysis, it can be concluded that the surface roughness impacts the metastable pit formation, the smaller the surface roughness, the less formation of metastable pits. The values of the passive current density go hand in hand with those of the E break . The smaller the anodic current, (at E break ), a denser passive oxide film is formed. Therefore, a smaller current density is preferred. All of the samples had a current density of around 0.4-3 µA/cm 2 , except for the PC, which had a larger current density of around 12 µA/cm 2 , at E break .
The E prot is the point of intersection of the forward and reverse scans. The higher the E prot , the least prone to corrosion the material is [56]. Therefore, from Table 7, the smallest E prot was identified for the PSPC sample, while the largest was that of the wrought. This shows that the wrought has a stable passive film and pit growth is restricted at an earlier potential value.    All of the six curves show a positive hysteresis loop, which is linked to pitting. The larger the hysteresis loop, the more location for pitting to occur and the less pitting corrosion resistant the material is. Therefore, the bigger the hysteresis loop, the more damage that is occurring on the passive oxide film and the more difficulty to restore it. All the curves show a positive hysteresis loop, with the biggest hysteresis loop identified for the PP. This was determined by finding the difference between E break and E prot , as shown in Table 7. Figure 15a shows the three curves with similar behaviour, that of the W, AP and SP. The results in Figure 15b show that the PC had a poor corrosion resistance when comparing it to the PP and the PSPC, which performed the best. The testing solution damaged the coating and formulated pits which reach the substrate, damaging the sample. As already mentioned, the surface roughness plays an important part in the formation of pits. The PC has a surface roughness of 8 ± 1 µm, which is similar to the 10 ± 3 µm of the AP. This shows that the higher the surface roughness the more surface area for pits to form.
Surface Analysis after Testing Figure 16 shows SEM micrographs of each sample after corrosion testing. Severe damage to the coupons was not evident. However, pits and pores of different sizes were visible on all the samples. In the PC micrographs, the printing striations and directions were revealed, while the PSP and the PSPC samples both show the dimples on the surface which are a characteristic of SP. Additionally, in the PSPC sample, a part of the TiAlCuN coating was delaminated following corrosion testing. Figure 17a,b shows that, upon inspecting the PC sample at a higher magnification, multiple pits were found, as opposed to those found on the shot peened ( Figure 17c) and hybrid treated (Figure 17d). On the PSPC less pits were identified, whose benefit will be explained later. The above demonstrates that the PC has the most corrosion susceptibility, while the PSPC has the least corrosion susceptibility, indicating that the combined effect of the surface treatments of shot peening and PVD provided superior corrosion qualities.

Conclusions
This study was carried out to analyse the effect of the surface treatments of shot peening and TiAlCuN coating on AP additive manufactured 316L SS, on the surface and sub-surface of the material. The main conclusions from this study include:

•
Microscopy and XRD phase analysis showed that the as-printed 316L SS was composed of an austenitic matrix, characterised with columnar and cellular dendritic together, together with the presence of some ferrite. • XRD stress measurement highlighted tensile residual stresses in the as-printed samples and compressive residual stresses in the shot peened and hybrid treated samples. Compressive residual stresses of 589 MPa for an approximate depth of 250 μm were generated by the cold working achieved by shot peening. • A 40% increase in surface hardness was obtained on the printed and shot peened specimens, while a 2.9 times increase was achieved following the application of the combined surface treatments. This shows that the coating possesses a high hardness, which when combined with shot peening, improves the material characteristics. The above demonstrates that the PC has the most corrosion susceptibility, while the PSPC has the least corrosion susceptibility, indicating that the combined effect of the surface treatments of shot peening and PVD provided superior corrosion qualities.

Conclusions
This study was carried out to analyse the effect of the surface treatments of shot peening and TiAlCuN coating on AP additive manufactured 316L SS, on the surface and sub-surface of the material. The main conclusions from this study include:

•
Microscopy and XRD phase analysis showed that the as-printed 316L SS was composed of an austenitic matrix, characterised with columnar and cellular dendritic together, together with the presence of some ferrite. • XRD stress measurement highlighted tensile residual stresses in the as-printed samples and compressive residual stresses in the shot peened and hybrid treated samples. Compressive residual stresses of 589 MPa for an approximate depth of 250 µm were generated by the cold working achieved by shot peening. • A 40% increase in surface hardness was obtained on the printed and shot peened specimens, while a 2.9 times increase was achieved following the application of the combined surface treatments. This shows that the coating possesses a high hardness, which when combined with shot peening, improves the material characteristics.
• A 50% decrease for R a and an 80% decrease for R z were found following the application of shot peening on the as-printed specimens. This shows that shot peening has the added advantage of improving the surface finish of additive manufactured components.

•
The application of the TiAlCuN coating on the as-printed provided better adhesion characteristics of the additive manufactured 316L stainless steel, than on the hybrid counterpart. This could be attributed to the test mechanics, where the tip interaction is changing with the roughness and the improved load support provided by the harder and stiffer coating.

•
The printed and coated combination had the worst corrosion behaviour, while the printed and hybrid treated specimens exhibited the best corrosion behaviour showing that the combined effect of the surface treatments of shot peening and PVD provided optimal corrosion qualities.
The results achieved in this study show optimal qualities for applying a shot peening treatment combined with the deposition of a coating on additive manufactured 316L stainless steel, making this combination of material processing ideal for a range of demanding applications involving bulk mechanical loading, susceptibility to wear under contact loads and corrosion damage, including many such instances found in the maritime transportation industry.