Evaluation of nano mechanical properties and corrosion behavior of diffusion welded CoCrNi medium-entropy alloy (MEA) and SUS 304 stainless steel joints

2.1. Preparation of CoCrNi (MEA)

A vacuum diffusion bonding furnace (FJK-2, China) was used for joining the CoCrNi MEA and SUS 304 alloys. A vacuum arc-melting furnace (VAM) under a high purity argon gas atmosphere with water-cooled copper mold was used to melt and cast the CoCrNi MEA alloy. A batch of 100 g with an equiatomic ratio was placed in the mold such that the element with the highest melting point (Cr) was placed on the top or nearest to the electric arc and the low melting element (Ni) was positioned at the bottom to avoid metal evaporation. The solidified button- ingots were re-melted five times to avoid chemical segregation. To confirm the chemical homogeneity of CoCrNi MEA, elemental mapping was performed using SEM. Figure 1a shows the sample location for elemental mapping and the corresponding SEM scans of as prepared CoCrNi MEA are depicted in Fig. 1(d and e). Macroscopic views of the sample were also presented in Fig. 1(a-c). Results displayed the uniform distribution of all the constituent elements (i.e. Co, Cr, and Ni). Therefore, no evidence for the development of intermetallic compounds was seen which is also consistent with the literature (Wang et al., 2019Wang, J., Yang, H., Ruan, J., Wang, Y., Ji, S. (2019). Microstructure and properties of CoCrNi medium-entropy alloy produced by gas atomization and spark plasma sintering. J. Mater. Res. 34 (12), 2126-2136. https://doi.org/10.1557/jmr.2019.96.; Samiuddin et al., 2021Samiuddin, M., Li, J., Chandio, A.D., Muzamil, M., Siddiqui, S.U., Xiong, J. (2021). Diffusion welding of CoCrNi medium entropy alloy (MEA) and SUS 304 stainless steel at different bonding temperatures. Weld. World 65 (11), 2193-2206. https://doi.org/10.1007/s40194-021-01165-5.; Zhu et al., 2021Zhu, M., He, F., Yuan, Y., Guo, S., Wei, G. (2021). A comparative study on the corrosion behavior of CoCrNi medium-entropy alloy and 316L stainless steel in simulated marine environment. Intermetallics 139, 107370. https://doi.org/10.1016/j.intermet.2021.107370.).

2.2. Diffusion welding of CoCrNi MEA and SUS 304 joints

Before diffusion bonding, the sample’s surfaces were grounded and cleaned with acetone for 5 min. After that these samples were assembled in a furnace between the jigs for the bonding process (as shown in Fig. 2a). Samples with faying surfaces parallel to each other were placed in jigs with a thin mica layer between bonded couple material and jig to evade adhesiveness. The alloys were bonded using three different temperatures (i.e. 950 °C, 1000°C, and 1050 °C) at 1 h bonding time and 5 MPa pressure with a vacuum level of 10^-3 Pa. The welding cycle and other parameters used in the diffusion welding process are shown in Fig. 2b.

2.3. Electrochemical analysis

The corrosion behavior of the diffusion welded samples were studied using a computer-based three-electrode system (Potentiostat G750, Gamry Instruments, USA). The samples were prepared by soldering copper wire at one side followed by epoxy mounting. The samples were ground up to 600 mesh by using silicon carbide paper and polished in alumina suspension. The polished samples were sonicated in ethanol for 30 min to remove surface contamination. The samples were further washed with distilled water and dried with hot air. An area of about 0.3 cm² in circular shape was used in such a manner that the bond interface sat along its diameter, enabling a portion of both the alloy exposed to 0.5 M HCl solution. The prepared sample was immediately immersed in the three-electrode electrochemical cell having a one-liter de-aerated solution of 0.5M HCl. The solution was prepared by using 36% pure hydrochloric acid of SigmaAlched® and the prepared solution was de-aerated by purging argon gas for 30 min from the solution. In the electrochemical cell, the counter electrode was graphite while the reference electrode was Saturated Calomel Electrode (SCE). For electrochemical impedance spectroscopy (EIS) analysis, an open-circuit delay of one hour was given to all samples and the test was conducted in a frequency range of 100 kHz to 0.01 Hz in compliance with ASTM G3-89 and G5-94 standards. For potentiodynamic polarization analysis, the samples were polarized in both cathodic and anodic directions in a potential difference range of ± 0.5 volts vs open-circuit potential. An optical microscope was used to observe the corroded surface after the potentiodynamic polarization test. All the tests were conducted at room temperature in the steady-state conditions and an average of three test results were considered in establishing the results.

2.4. Instrumented nanoindentation tests

Instrumented nanoindentation test was executed on a polished surface of diffusion welded samples using Anton Paar (NHT³, USA) in compliance with ASTM E2546-07 standard. Indenter of Berkovich type (B-U 55) was employed at 50mN/min loading/unloading rate and a holding time of 15 sec was used to eradicate the time-dependent creep response of the specimens. Five different loads ranging from 20 mN to 100 mN (with a 20 mN difference) were applied to determine the nanomechanical properties across the weld interface. After the test, all the specimens were observed using an optical microscope to confirm the correct indenter’s position and its impression. Averages of three readings were considered in establishing the results.

3.1. Microstructural examination of the weld interface

After metallographic preparation, all the samples were etched through aqua regia (15 ml HCl, 5 ml HNO₃) and subsequently, SEM (Tescan Mira 3, The Czech Republic) was used to characterize the weld interface, as depicted in Fig. 3. It can be seen in Fig. 3a that the weld interface formed at 950 °C exhibited the formation of intermetallic compounds (IMCs). Detailed discussion on the nature of these IMCs and their formation can be found in (Muhammad Samiuddin et al., 2021Muhammad Samiuddin., Jinglong, L., Siddiqui, M.A., Jiangtao, X., Ling, R. (2021). Electrochemical Corrosion Behavior of CoCrNi Medium Entropy Alloy and SUS-304 Stainless Steel Diffusion Welded Joints at Various Bonding Temperatures. Phys. Metals Metallogr. 122 (14), 1561-1571. https://doi.org/10.1134/S0031918X21140234.; Samiuddin et al., 2021Samiuddin, M., Li, J., Chandio, A.D., Muzamil, M., Siddiqui, S.U., Xiong, J. (2021). Diffusion welding of CoCrNi medium entropy alloy (MEA) and SUS 304 stainless steel at different bonding temperatures. Weld. World 65 (11), 2193-2206. https://doi.org/10.1007/s40194-021-01165-5.). In the light of EDX point analysis, it is suggested that most of the IMCs were Cr-rich carbides as shown in Fig 3d. Similar results were also reported in several publications (Muhammad Samiuddin et al., 2021Muhammad Samiuddin., Jinglong, L., Siddiqui, M.A., Jiangtao, X., Ling, R. (2021). Electrochemical Corrosion Behavior of CoCrNi Medium Entropy Alloy and SUS-304 Stainless Steel Diffusion Welded Joints at Various Bonding Temperatures. Phys. Metals Metallogr. 122 (14), 1561-1571. https://doi.org/10.1134/S0031918X21140234.; Samiuddin et al., 2021Samiuddin, M., Li, J., Chandio, A.D., Muzamil, M., Siddiqui, S.U., Xiong, J. (2021). Diffusion welding of CoCrNi medium entropy alloy (MEA) and SUS 304 stainless steel at different bonding temperatures. Weld. World 65 (11), 2193-2206. https://doi.org/10.1007/s40194-021-01165-5.).

The severity of the formation of such IMCs along with the weld interface drastically reduced at 1000 °C bonding temperature (see Fig 3b). Thus, with the increase of bonding temperature suppression of IMCs formation occurred which is due to the less thermal stability of Cr-rich carbides above 1000 °C (Kleykamp, 2001Kleykamp, H. (2001). Thermodynamic studies on chromium carbides by the electromotive force (emf) method. J. Alloys Compd. 321 (1), 138-145. https://doi.org/10.1016/S0925-8388(01)01012-X.; Kashaev et al., 2018Kashaev, N., Ventzke, V., Stepanov, N., Shaysultanov, D., Sanin, V., Zherebtsov, S. (2018). Laser beam welding of a CoCrFeNiMn-type high entropy alloy produced by self-propagating high-temperature synthesis. Intermetallics 96, 63-71. https://doi.org/10.1016/j.intermet.2018.02.014.). Consequently, a negligible amount of IMCs was present along the weld interface formed at 1050 °C bonding temperature, as shown in Fig. 3c. Furthermore, EDX line scanning was utilized to measure the bond interface of the welded joints and representative profiles are displayed in Fig. 3(e and f).

In order to determine the effect of processing parameters on IMCs, its quantitative analysis were also performed using image analysis software. It has been discovered that the size and amount of IMCs decrease over time and with increasing bonding temperature as shown in Table 1. It appeared that the IMCs gradually dissolved during the bonding process, and the rate of dissolution accelerated at higher bonding temperatures. Ultimately, these changes in IMC characteristics have implications on the interface properties. At 950 °C smaller size of IMCs (up to 489 nm) were formed at 1 h bonding time which significantly grew (up to 900 nm) at 3 h bonding time almost double comparative to 1 h bonding time. While at 1000 °C and 1050 °C the IMCs size were further grown with a difference of less than 100 nm within the time span of 1 h to 3 h which indicated that bonding time did not significantly influenced the growth rate of the IMCs. Similar effect of the bonding time was also perceived at 1050 °C at different time span. Although, the area fraction of the IMCs was significantly pretentious with the change of both the parameters. The area fraction of the IMCs was reduced to ~35 % (with 1 h) from ~48 % (with 3 h) at 950 °C, while it further decreased to ~5 % at 1000 °C with 3 h bonding time as depicted in Table 1 and almost shrink to ~0.42% at 1050 °C and 3 h bonding time. This strongly suggested that the IMCs became unstable at higher parametric values and dissolved along the interface due to the paramount effect of IMCs thermodynamics, constituent’s atomic diffusion, and their interaction with IMCs, and the concentration gradients across the interface etc.

3.2. Nanoindentation behavior of weld interface

Nano-mechanical properties across the weld interface of diffusion welded samples under different indentation loads were studied using the nanoindentation technique. Parametric loading values were selected in such a way that the corresponding indent size was not exceeded by the width of the weld interface. Thus, loading values of 20 mN to 100 mN were selected which corresponds to less than 9.86 µm maximum indent size. Thereby, the nanoindentation behavior of only weld interfaces under different bonding temperatures was evaluated. A matrix of indents across the weld interface was formed on a well-polished surface of diffusion welded coupons as depicted in Fig. 4. The manner in which indents were taken is shown in Fig. 4d.

A pronounced effect of grain coarsening on the SUS 304 side can be seen on the samples welded at 1050 °C in contrast with other samples (see Figs. 4 (a-c)). However, there were no such appreciable changes in grain size of CoCrNi MEA transpired due to its sluggish diffusion characteristics, attributed to HEAs. Weld interface of 14.25 µm was formed at 1050 °C bonding temperature which was relatively thick compared to 950 °C (9.78 µm) temperature as shown in Fig. 3(e and f). Thus, the thickness of the weld interface was also built up with the increase in welding temperature. Optical images of SUS 304 and CoCrNi MEA are also shown in Fig. 4(e and f) respectively. CoCrNi MEA was etched with acetic glyceregia with the concentration of 15 ml HCL, 10 ml acetic acid, 5 ml HNO3, and 1-2 drops of glycerol, whereas, standard Aqua regia was used for etching SUS 304 stainless steel. It is observed that relatively coarser grains (up to 150 µm) were formed in MEA while SUS 304 stainless steel exhibited a grain size of 25 µm.

Figure 5 depicted load-displacement curves plotted against different nanoindentation loadings using Anton Paar TriTec version 6.0x software. Figure 5a indicated scans of nanoindentation across the weld interface of the 950 °C bonded sample. Significant variations in the slope of the loading curve can be easily grasped. The loading curves became less stepper with the increase of indentation loads which indicated a low deformation rate. Since 950 °C bonded sample contained lots of brittle IMCs along with its weld interface (see Fig. 3a) which imposed hindrance on deformation. Since at low indentation values the character of IMCs became more prominent due to comparable indent size to IMCs. Similarly, upon increasing indentation load the indent size widened and facilitates deformation due to the combined effect of IMCs and their surrounded region (matrix). The discrepancies in deformation rate upon increasing indentation load became less ascetic on sample welded at 1000 °C bonding temperature as its weld interface contained fewer IMCs compared with 950 °C bonded sample. More IMCs from the weld interface were eradicated with a further increase in welding temperature (i.e. 1050 °C) which subsequently promoted a uniform loading rate on all indentation loads as represented in Fig. 5c.

3.3. Nanomechanical properties of the weld interface

Figure 6 displayed nanomechanical properties extracted from the load-displacement curves (Fig. 5) using the approach proposed by Oliver and Pharr (2004)Oliver, W.C., Pharr, G.M. (2004). Measurement of hardness and elastic modulus by instrumented indentation: Advances in understanding and refinements to methodology. J. Mater. Res. 19 (1), 3-20. https://doi.org/10.1557/jmr.2004.19.1.3.. The nano hardness of SUS 304 steel was found in the range of 4.4 + 2 GPa with an elastic modulus of 225 + 30 GPa under different loads. While CoCrNi MEA exhibited nano hardness and elastic modulus of 3.98 + 1.2 GPa and 290 + 18 GPa respectively. Nano hardness across the weld interface was also mapped by taking the average of three indentations at each load as shown in Fig. 4. Since the nature of weld interface changes with the dissolution of IMCs in response to welding temperature. As a result nanomechanical properties across the weld interface also varied. Figure 6 revealed the variation of nano hardness and elastic moduli as the indentation load and bonding temperature changed. Fig. 6a demonstrates a declining trend of nano hardness with bonding temperature. The maximum value of nano hardness (i.e. 3.82 GPa) was achieved with a sample welded at 950 °C bonding temperature signifying the presence of IMCs. When bonding temperature increases a substantial drop in nano hardness (i.e. 1.58 GPa) was observed associated with carbide dissolution as discussed previously. Another reason that can be attributed to the variation of nano hardness is related to the indentation size effect (ISE). According to the conventional theory of ISE, nano hardness increases with the reduction of indent size due to the dislocation forest hardening mechanism (Voyiadjis and Yaghoobi, 2017Voyiadjis, G.Z., Yaghoobi, M. (2017). Review of Nanoindentation Size Effect: Experiments and Atomistic Simulation. Crystals 7 (10), 321. https://doi.org/10.3390/cryst7100321.). It can be seen that there was a significant increase of nano hardness at low indentation load (i.e. 20 mN) and it progressively drop until at 40 mN load, after which it became stable. This specified that the indentation size effect (ISE) prevailed below 40 mN indentation load also shown in Fig. 6d.

Elastic moduli of all the samples were also calculated from nanoindentation data and the results were plotted in Fig. 6c and d, against various bonding temperatures and indentation loads. Elastic moduli were also influenced by the bonding temperature, a similar decreasing trend in elastic moduli was also observed as in the case with nano hardness. This is due to the presence of IMCs which hindered deformation against the indentation, consequently, both elastic moduli and hardness increase. However, an insignificant change in elastic moduli was observed when different indentation loads were applied as shown in Fig. 6d. This indicated that elastic moduli are insensitive to the change of indentation load.

Figure 7 represented the elastic properties of the weld interface evaluated through a nanoindentation test. It was noticed that indentation depth (H_final) increases with the rise of bonding temperature which also reinforced the fact that IMCs became unstable and ultimately dissolved at the high bonding temperature. Thus, with the diminution of IMCs weld interface became more elastic and experienced large deformation. Consequently, higher values of indentation depth (H_final) was achieved with increasing indentation load, as depicted in Fig. 7a. Additionally, it was also revealed from Fig. 5 that upon removal of load, strain retrieval transpired from the weld interface resulted in a reduction of H_final values with the increase of indentation load. Therefore, elastic recovery offered by the weld interface can be determined from the unloading path by taking the difference between the two indentation depths (i.e. H_max & H_final) defined in Fig. 5d (Muthupandi et al., 2017Muthupandi, G., Lim, K.R., Na, Y.-S., Park, J., Lee, D., Kim, H., Park, S., Choi, Y.S. (2017). Pile-up and sink-in nanoindentation behaviors in AlCoCrFeNi multi-phase high entropy alloy. Mater. Sci. Eng. A 696, 146-154. https://doi.org/10.1016/j.msea.2017.04.045.). Figure 7b displayed elastic recovery of the weld interface which is also compliant with previous observations. Thus a high yield of elastic recovery is achieved at high bonding temperature due to the elimination of carbides as these carbides offered resistance to deformation (Hussainova et al., 2008Hussainova, I., Jasiuk, I., Du, X., Cabassa, D. Pirso, J. (2008). Mechanical properties of chromium carbide based cermets at micro - level. In Advances in powder metallurgy & particulate materials. PowderMet 2008, Washington, DC, USA.). Similarly, the weld interface with lesser IMCs allowed more deformation upon higher indentation load which resulted in higher elastic recovery as shown in Fig. 7b.

In summary, the formation of IMCs influences the deformation rate during the indentation process as indicated in Fig. 5. Changes in the deformation rate can be visually perceived from the variation of slope in the loading curves which was higher in the sample welded at 950 °C as shown in Fig. 5a. Such variation in the slope of loading curves was drastically reduced as the welding temperature increased which might be due to the suppression of IMCs formation (see Fig. 5c). Furthermore, regarding the formation of these brittle IMCs, it was evident from the hardness data that an upshot in Nano hardness was present in the sample welded at 950 °C while it significantly drop with the suppression in IMCs formation as shown in Fig. 6a. However, ISE also prevailed and such effect was impassive when the load was kept higher than 40 mN (see Fig. 6b). Therefore, 60 to 100 mN were the optimized parameter through which the nano-mechanical properties of the welded joints can be determined with certainty. Elastic moduli were also influenced by the presence of such IMCs and it is shown that decreasing trend in the Elastic moduli was observed with the increase in welding temperature (see Fig. 6c). This indicated that the presence of IMCs tends to lower the deformation rate which ultimately increases the Elastic moduli values. Moreover, a significant increase in elastic recovery and h_final also represented that these IMCs were brittle, and removal of which induced softness in the weld interface and consequently elastic recovery increases.

3.4. Corrosion behavior of weld interface

This section deals with the corrosion performance of the diffusion welded samples under 0.5M HCl environment. Figure 8 showed the potentiodynamic polarization curves, the pitting potential, and the I_corr values of all samples along with their comparison with SUS 304 and CoCrNi MEA. It is observed from Fig. 8a that the E_corr value of 304SS (-270 mV) was more negative to CoCrNi alloy (-194 mV) in the given solution. For this reason, SUS 304 acted as an anode when welded with CoCrNi alloy due to the formation of galvanic cells. Also, the corrosion due to galvanic coupling was not much severe as the difference in electrode potential was not significantly high (76 mV) and the coupling was favorable. From Fig. 8b, it is found that E_corr of diffusion welded samples were in between the potential value of SUS 304 and CoCrNi MEA. Moreover, the sample diffusion welded at 1000 ºC temperature was most electropositive followed by the sample welded at 950 ºC and 1050 ºC bonding temperature. Figure 8c revealed that the pitting potential of CoCrNi MEA (336 mV) is much higher than SUS 304 (6 mV). After diffusion welding, the pitting potential was significantly dropped as compared to CoCrNi MEA. Among the diffusion welded samples, the sample welded at 1000 ºC grasped the highest pitting potential (192 mV) value followed by the sample welded at 950 ºC (172 mV), while the sample welded at 1050 ºC bonding temperature exhibited the lowest pitting potential (120 mV). The corrosion current density (I_corr) of all the samples was also calculated from the Tafel analysis and displayed in Fig. 8d. It is found that the I_corr value of CoCrNi alloy is much smaller than SUS 304 while among diffusion welded samples, the sample welded at 1000 ºC temperature showed the lowest I_corr value; even smaller than CoCrNi MEA. The remaining two diffusion welded samples exhibited the similar I_corr value. The corrosion-resistant properties of SUS 304 were found inferior in comparison to CoCrNi alloy because the concentration of corrosion-resistant alloying elements (i.e. Cr, Ni, etc.), was higher in CoCrNi MEA as compared to SUS 304 (Kumar et al., 2017Kumar, N., Fusco, M., Komarasamy, M., Mishra, R.S., Bourham, M., Murty, K.L. (2017). Understanding effect of 3.5 wt.% NaCl on the corrosion of Al0.1CoCrFeNi high-entropy alloy. J. Nucl. Mater. 495, 154-163. https://doi.org/10.1016/j.jnucmat.2017.08.015.; Luo et al., 2018Luo, H., Li, Z., Mingers, A.M., Raabe, D. (2018). Corrosion behavior of an equiatomic CoCrFeMnNi high-entropy alloy compared with 304 stainless steel in sulfuric acid solution. Corros. Sci. 134, 131-139. https://doi.org/10.1016/j.corsci.2018.02.031.).

The Nyquist plot, displayed in Fig. 9 compared SUS 304, CoCrNi MEA, and diffusion welded samples at various bonding temperatures. In this plot, the diameter of the half-circle represented the polarization resistance of the samples. This shows that greater the diameter of the half-circle, the higher will be the polarization resistance of the material. Thus, in this regard, the polarization resistance of the sample welded at 1000 ºC is found higher followed by the sample welded at 950 ºC, while the polarization resistance of the sample welded at 1050 ºC was most inferior among all the welded samples. Furthermore, it is also perceived that after diffusion welding polarization resistance increases because the performance of both the alloys before welding was inferior as observed in Fig. 9.

Table 1 shows the data obtained by fitting Randell modified equivalent circuit as shown in Fig. 9. In this circuit, “Rs” is solution resistance which is attached in series with “Rp”; a polarization resistance and “ Yo”; a constant phase element (CPE) magnitude, and “n” is the CPE exponent. The pure capacitor has been replaced by CPE to minimize the effect of surface irregularities produced during sample preparation and other inherent surface microstructural defects of the material. Double-layer capacitance was calculated by using Eq. (1). Generally, the capacitance has an inverse relation with surface film thickness or in other words, passive layer thickness. The passive layer thickness was calculated by using Eq. (2) and Eq. (3) (Wallinder et al., 1998Wallinder, D., Pan, J., Leygraf, C., Delblanc-Bauer, A. (1998). EIS and XPS study of surface modification of 316LVM stainless steel after passivation. Corros. Sci. 41 (2), 275-289. https://doi.org/10.1016/S0010-938X(98)00122-X.; Sorour et al., 2021Sorour, A.A., Farooq, M., Mekki, A., Kumar, A.M. (2021). Corrosion of a Spark Plasma Sintered Fe-Cr-Mo-BC Alloy in Hydrochloric Acid. Metallogr. Microstruct. Anal. 10 (3), 291-301. https://doi.org/10.1007/s13632-021-00739-5.).

In equation (1), “ω_max” is an angular frequency (ω = 2πf ) in rad/sec corresponding to the maximum “-Z_img” value (Machnikova et al., 2008Machnikova, E., Whitmire, K.H., Hackerman, N. (2008). Corrosion inhibition of carbon steel in hydrochloric acid by furan derivatives. Electrochim. Acta 53 (20), 6024-6032. https://doi.org/10.1016/j.electacta.2008.03.021.) which was obtained directly from experimental data. Based on “n” value, constant phase element (CPE) can be classified as pure resistance (R) when (n=0), Warburg (W) when (n=0.5) and inductance (L) when (n = -1). CPE was considered as pure double-layer ideal capacitance (C_dl) when the value of n = 1 (Behpour et al., 2009Behpour, M., Ghoreishi, S.M., Soltani, N., Salavati-Niasari, M. (2009). The inhibitive effect of some bis-N, S-bidentate Schiff bases on corrosion behaviour of 304 stainless steel in hydrochloric acid solution. Corros. Sci. 51 (5), 1073-1082. https://doi.org/10.1016/j.corsci.2009.02.011.). In this case, all samples exhibited both resistive and capacitive behavior. InEq.(2), “C” is double-layer capacitance, “-Z_img” and “ω” is an imaginary impedance and angular frequency respectively, corresponding to 1 kHz frequency. In equation (3), “d_ox” is passive film thickness, “ԑ” is the permittivity of the dielectric layer for chromium oxide (Cr₂O₃) which was considered as “12” and “As” is the exposed surface area of the working electrode. Since the sample was immersed in a conductive solution, the permittivity of free space “ε_o” can be omitted (McCafferty, 2010McCafferty, E. (2010). Introduction to corrosion science. Springer Science & Business Media.).

From Table 2, it is observed that the polarization resistance of CoCrNi alloy was greater than 304 SUS while the passive film thickness of SUS 304 is slightly greater than CoCrNi MEA. It is also observed that the polarization resistance (R_p) and passive film thickness of diffusion welded samples were greater in comparison to SUS 304 and CoCrNi MEA. Moreover, among diffusion welded samples, it was observed that the polarization resistance and passive film thickness of the sample welded at 1000 ºC were highest as compared with 950 ºC and 1050 ºC. Thus, both R_p and passive film thickness are augmented as bonding temperature increases. From the above discussion, it can be concluded that the optimum bonding temperature concerning polarization resistance and passive film thickness is 1000 ºC.

Surface examination after potentiodynamic scanning was also executed on diffusion welded samples. For this purpose, optical microscopy was utilized and the results are displayed in Fig. 10. It is seen that at 950 °C bonding temperature the severity of pit formation was relatively greater. Large patches of materials (pits) up to ~443 µm were removed from SUS 304 steel side while the MEA side remained unaffected which implied that CoCrNi offered excellent resistance against pit formation which is in well agreement with observed results in the potentiodynamic polarization scan. Moreover, corrosion was also anticipated along the weld interface resulting in a thick and prominent dark line as shown in Fig. 10a, with significant removal of material along with the weld interface. As the bonding temperature increased to 1000 °C, the instigation of corrosion shifted towards SUS 304 stainless steel and limited dissolution along the weld interface ensued. Consequently, a significant reduction in pit formation occurred and pit size up to ~78 µm was perceived, as displayed in Fig. 10b. With 1050 °C bonding temperature, the sample contained a pit size up to ~203 µm indicating the corrosion resistance was in between the two welded samples. Moreover, owing to the high bonding temperature, the sample also exhibited intensified intergranular corrosion within SUS 304 side while pit formation ~169 µm was also initiated at the MEA side as shown in Fig. 10c.

Variation in corrosion resistance of diffusion welded samples joined at different bonding temperatures can be explained based upon IMCs formation. Due to the different electrochemical properties of IMCs and the matrix, localized corrosion along the weld interface is instigated in response to the micro galvanic effect between IMCs and the matrix (Silva et al., 2020Silva, D.D.S., Lima, L., Araujo, A.J.M., Silva, V.D., Raimundo, R.A., Damasceno, I. Z., Simoes, T.A., Gomes, R.M. (2020). The Effect of Microstructural Changes on Mechanical and Electrochemical Corrosion Properties of Duplex Stainless Steel Aged for Short Periods. Materials 13 (23), 5511. https://doi.org/10.3390/ma13235511.; Silva et al., 2021Silva, D.D.S., Simões, T.A., Macedo, D.A., Bueno, A.H.S., Torres, S.M., Gomes, R.M. (2021). Microstructural influence of sigma phase on pitting corrosion behavior of duplex stainless steel/NaCl electrolyte couple. Mater. Chem. Phys. 259, 124056. https://doi.org/10.1016/j.matchemphys.2020.124056.). Thus, corrosion resistance increased when the formation of IMCs along the weld interface was inhibited. Although the weld interface at 1050 °C bonding temperature was almost free from IMCs despite this it exhibited poor corrosion resistance. Since at high bonding temperature more Cr is deposited along the grain boundaries (see thick grain boundaries in Fig. 3c, 4c, and 10c) leading to the intergranular corrosion at SUS 304 side. Results obtained from potentiodynamic polarization scanning also supported the observations extracted from optical imaging.