A. Base metals microstructures

The microstructure of the Hastelloy C-276 BM consisted of an austenitic equiaxed grains as shown in Fig. 1(a). On the other hand, BM of Type 321 ASS consisted of fine austenitic grains along with small cuboidal inclusions of titanium carbides as shown in Figs. 1(b) and 1(c), respectively.

FIG. 1.

Optical micrographs of BMs; (a) Hastelloy C-276; (b) Type 321 ASS; (c) EDS spectrum showing presence of cuboidal shape Ti-rich particles; (d) image showing macrograph of dissimilar welded Hastelloy C-276 and Type 321 ASS.

B. Macrostructural examination

Macrostructural examination was carried out on the dissimilar joint fabricated by GTAW as depicted in Fig. 1(d). The macrograph showed proper fusion and good weld bead profile, which construed that the parameters chosen and the selection of filler material were suitable with the BMs.

C. Microstructural examination of the dissimilar weldments

The fusion zone (FZ) microstructure to a great extent depends on the cooling rate and has a significant influence on the mechanical properties of welded joints. ^{Reference Dupont, Banovic and Marder5} Fig. 2 showed that the FZ microstructures obtained upon solidification were fully austenitic due to the presence of austenite promoting elements (such as Ni, Mo, and Co). ^{Reference Baghjari and Mousavi19} However, the morphology of the subgrain solidified structure differed across the various passes [locations shown in Fig. 2(a)]. The subgrain solidification structure in the root pass and the cap pass mostly consisted of a combination of columnar and equiaxed morphology [Figs. 2(b) and 2(e)] as compared to coarser columnar dendritic structure appeared in filler pass-1 and filler pass-2 [Figs. 2(c) and 2(d)]. As reported in the literature, the tendency of the structure to change from the columnar dendritic structure to equiaxed dendritic structure increases with increasing cooling rate. ^{Reference Kuo and Lee20} The higher temperature gradients at the root and cap passes resulted in faster cooling rate and hence induced variation in the solidification dendritic structure. ^{Reference Ramkumar, Anirudh, Singh, Goyal, Gupta, George and Arivazhagan21}

FIG. 2.

(a) Schematic view of FZ containing different passes; optical microstructures of (b) 1st: root pass (c) 2nd: filler pass-2 (d) 3rd: filler pass-3 (e) 4th: cap pass; (f, g) weld interfaces on Type 321 ASS side and Hastelloy C-276 side, respectively.

Figures 2(f) and 2(g) shows the optical microstructures at the interface of dissimilar welded Type 321 ASS and Hastelloy C-276 joints. Owing to the similarity in the crystal structures of both the BMs and FZ, crystal orientations of the BM influence the microstructure of the weld and lead to form the epitaxial growth region. ^{Reference Baghjari and Mousavi19} The interfacial microstructures showed a well-defined fusion line along with dendritic growth at the weld interface on both Type 321 ASS and Hastelloy C-276 side [locations mentioned as (i) and (ii) in Fig. 2(a) respectively]. Figure 2(f) showed the presence of an unmixed zone (UMZ) at the interface on Type 321 ASS side. This occurs when the melting point of filler metal (≈1370 °C) is similar/higher to that of BM (≈1400 °C), then only a small fraction of the BMs can be melted and no dilution takes place in the re-solidification stage which ends up in the formation of an UMZ. ^{Reference Sayiram and Arivazhagan3} Furthermore, the presence of δ-ferrite stringers was also observed at the heat affected zone (HAZ), which has a beneficial effect of minimizing grain growth and decreasing propensity to HAZ liquation cracking. ^{Reference Kumar, Ramkumar and Arivazhagan22} On the other hand, the microstructures at the weld interface of Hastelloy C-276 side had undergone grain coarsening in the HAZ and also divulged small presence of a partially melted zone (PMZ) characterized by grain boundary melting and thickening as shown in Fig. 2(g). The generation of higher heat input during GTAW might have caused the formation of PMZ and also to the weld pool convection which causes separation of the partially melted grains from the solid–liquid (S–L) mixture surrounding the weld pool. ^{Reference Ramkumar, Naren, Paga, Tiwari and Arivazhagan23}

D. SEM/EDS analysis of the dissimilar weldments

The amount of segregation occurred upon solidification often lead to the precipitation of harmful secondary phases which affects the mechanical properties as well as corrosion resistance of the welded joints. ^{Reference Dupont, Banovic and Marder5} It has been reported in the literature that a coarser columnar dendritic structure associated with conventional fusion welding (such as GTAW) processes mostly results in high degree of segregation and are more prone to the hot cracking phenomenon as compared to cellular/equiaxed dendritic structure. ^{Reference Hashim, Babu, Duraiselvam and Natu6} SEM-EDS point analysis was carried out across the weldment to evaluate the extent of microsegregation occurred within the dissimilar joints. The SEM image confirmed the presence of columnar dendrites in FZ as shown in Fig. 3(a). Further, EDS point analysis results showed that the subgrain bodies [Fig. 3(b)] were depleted in Mo and W and enriched in Ni and Fe as compared to subgrain boundaries [Fig. 3(c)]. This trend is believed to be based on the atomic size where elements of similar atomic radii mostly exhibit considerable solubility compared to elements having larger atomic radii. The size of both Fe (1.56 Å) and Cr (1.66 Å) is quite similar to matrix element Ni (1.49 Å) and hence do not segregate while Mo (1.90 Å) and W (1.93 Å) being larger in size (compare to Ni) segregate aggressively upon solidification of the weld. ^{Reference Dupont, Banovic and Marder5} The occurrence of white tiny particles occupying the subgrain boundaries of the columnar dendritic structure were also noticed in the FZ. These particles were highly enriched in Mo and W and depleted to a greater degree in Ni as shown in Fig. 3(d). As reported in the literature, it might be possible that some amount of intermetallic TCP phases such as sigma, P and μ could have formed during solidification of weld due to the slower cooling rate associated with GTAW process which needs further characterization. ^{Reference Manikandan, Arivazhagan, Rao and Reddy24}

FIG. 3.

SEM/EDS point analysis on the GTAW (a) FZ; (b–d) EDS analysis of (i) weld subgrain body; (ii) weld subgrain boundary; (iii) Mo rich phase; respectively.

Figures 4(a) and 4(b) revealed the presence of migrated grain boundaries (MGBs) at the interface on both Type 321 ASS side and Hastelloy C-276 side, respectively, which were not clearly visible in optical micrographs. MGBs are commonly prevalent in fully austenitic welds or due to the reheating actions occurred during multipass welding. ^{Reference Sayiram and Arivazhagan3} Figs. 4(c) and 4(d) showed the enlarged view of the interface region on both Type 321 ASS and Hastelloy C-276. EDS point analysis results depicted that the segregation of elements between weld interface subgrain boundaries and subgrain bodies were also noticed on both sides. Similar to FZ, the appearance of white tiny phases were also observed at the weld interface on Hastelloy C-276 side. However, the extent of segregation was comparatively higher at the weld interface of Hastelloy C-276 as compared to FZ as depicted in Fig. 4(e). Therefore, weld interface region was found to be more susceptible to hot cracking phenomenon as compared to FZ. ^{Reference Sayiram and Arivazhagan3} However, no sign of cracking was noticed in the present study which might be due to the proper selection of welding parameters.

FIG. 4.

(a, b) Images showing weld interfaces of Type 321 ASS and Hastelloy C-276 respectively; (c, d) enlarged view of interfaces on Type 321 ASS and Hastelloy C-276 side respectively along with EDS results; (e) EDS results showing precipitation of Mo-rich phase at the weld interface on Hastelloy C-276.

E. SEM/EDS analysis for compositional gradient across the composite region

Compositional gradient is considered as one of the important concerns during dissimilar joining of materials and leads to the movement of elements from higher concentration region to lower concentration region. The composite region which includes various zones [Type 321 ASS (BM + UMZ) + FZ (ERNiCrMo-4) + (PMZ + HAZ + BM) Hastelloy C-276] was subjected to SEM-EDS point analysis and results are shown in Fig. 5(a). The alloying elements (Fe, Cr, Mo, and Ni) which impose major influence on the properties of dissimilar weldments and also form a major part of the BMs matrix were chosen for EDS point analysis.

FIG. 5.

(a) SEM-EDS point analysis results across the weldment employing GTAW process; (b) horizontal microhardness plot from Type 321 BM to Hastelloy C-276 BM; (c) vertical microhardness plot from face passes to root passes.

A significant variation in the amount of Fe and Ni had observed to be migrated across Type 321 ASS/FZ interface. This enormous variation could be attributed mainly due to the huge concentration difference between Type 321 BM and filler wire. In addition, fast diffusion of Fe and Ni in the melt pool had also led to the large variation on account of small atomic radii as compared to other alloying elements. In addition, a slight movement of Mo from FZ toward Type 321 side was also observed. This might be due to the difference in the chemical composition of filler material (17%) and Type 321 BM (nil). However, variation in the Cr concentration was found to be almost negligible as both the BM and filler material consisted of an almost same amount. On the other hand, at the FZ/Hastelloy C-276 interface, lesser migration of element took place as the concentrations of these elements were almost similar. However, the concentration of Mo in the filler wire is 17%. From the EDS analysis, it was noticed that Mo has been depleted at the interface on Hastelloy C-276 side as compared to the center part of FZ and Hastelloy C-276 BM, as the percentage of Mo was found to be 12%. Formation of secondary phases [appeared as white in Fig. 4(e)] at the weld interface of Hastelloy C-276, had led to the depletion of this element. This might have happened due to the lower cooling rate coupled with GTAW process, which provides ample time for elements to get diffused and segregation to occur. In addition, when the solidification occurred in such a manner that the last solidifying liquid is predominantly rich in Mo and deprived in Ni, one could end up in the formation of secondary phase (s) and the same has happened at the weld interface of Hastelloy C-276. ^{Reference Manikandan, Arivazhagan, Rao and Reddy24}

F. Mechanical characterization

Microhardness measurement was carried out firstly in the longitudinal direction [location shown in Fig. 1(d)] that is on a plane perpendicular to the weld direction as shown in Fig. 5(b). It was observed that there was an increase in hardness as the indentor moves from Type 321 ASS BM side to Hastelloy C-276 BM side. The compositional difference associated with the BMs could be one of the reasons for such trend in microhardness. ^{Reference Hosseini, Shamanian and Kermanpur8} The average hardness at the FZ was found to be in between the hardness of BMs (281.6 HV) and could be attributed to the presence of columnar dendritic structure as reported by Sakthivel et al. ^{Reference Sakthivel, Vasudevan, Laha, Parameswaran, Chandravathi, Mathew and Bhaduri25} In addition, the microhardness increases suddenly at the fusion line on Hastelloy C-276 due to the presence of partially melted grains, which are partially adopted as nuclei by the new precipitating phase of the FZ during the solidification stage. ^{Reference Vashishtha, Taiwade, Khatirkar, Ingle and Dayal26} The variation in microhardness from cap pass to root pass is shown in Fig. 5(c). The results showed that microhardness near the top of the weld bead (cap pass) and also at the root region was higher as compared to other two passes. The reason for this could be attributed to the higher cooling rate at the surface of weld bead and root pass (the use of Cu back plates which promotes the cooling rate) which resulted in finer dendritic structure and consequently increased the hardness at both these passes. ^{Reference Kuo and Lee20}

Tensile test was conducted to ascertain the behavior of the weldment toward gradually increasing load. The rupture location propensity did not vary with increasing passes, so multi-pass welding could be an irrelevant factor in the tensile test. ^{Reference Kuo and Lee20} Further, the results showed that the fracture occurred at the weaker parent material side (i.e., Type 321 ASS BM) and tensile strength obtained was coming out to be 685.50 MPa as shown in Fig. 6(a). The hardness result was found to be in complete agreement with tensile results as Type 321 ASS BM experienced lower hardness. ^{Reference Ramkumar, Anirudh, Singh, Goyal, Gupta, George and Arivazhagan21} This also confirmed that the weld strength was quite higher compared to weaker parent material and the precipitation of Mo-rich phases in the FZ as well as the interface of Hastelloy C-276 did not weaken the matrix. ^{Reference Ramkumar, Anirudh, Singh, Goyal, Gupta, George and Arivazhagan21} As the failure occurred in the Type 321 ASS BM, it would be indecisive to comment on the role of FZ microstructure in case of TS. However, the presence of columnar in combination with equiaxed dendritic structure might have increased the restraint of FZ against failure. ^{Reference Sathiya, Mishra, Soundararajan and Shanmugarajan27} Fractured surface of the tensile specimen was observed to characterize the failure mode and the SEM fractograph is shown in Fig. 6(b). The SEM fractograph confirmed that the weldment failed due to ductile fracture by showing micro-voids/dimple appearance which coalesces together to undergo plastic deformation. ^{Reference Anand, Tamilmannan and Arivazhagan28}

FIG. 6.

Images showing (a) tensile specimen before and after failure; (b) SEM fractograph of tensile tested specimen; (c, d) images showing impact specimens before failure and after failure, respectively; (e) SEM fractograph of the impact tested specimen.

The images of impact specimens before and after the failure are shown in Figs. 6(c) and 6(d), respectively. The image clearly showed that the failure of specimen did not occur abruptly and had undergone sufficient deformation under impact loading. The average impact energy was found to be (182 J). This might be due to the fully austenitic structure which resulted in higher toughness of dissimilar welded joints. ^{Reference Manikandan, Arivazhagan, Rao and Reddy24} In addition, it has also been reported in the literature that as compared to columnar dendritic structure, equiaxed dendrites can accommodate higher strains and considered to be more ductile. ^{Reference Baghjari and Mousavi19} Therefore, the presence of equiaxed dendrites might also have contributed for increment in toughness value. The presence of fibrous network and micro-voids confirmed the ductile mode of failure as shown in Fig. 6(e).

G. Corrosion test

The dissimilar weldment was immersed in a “Green death” chloride solution to see the influence of corrosive environment on the tensile strength. The sample was removed after 24 h (at 100 °C) from the solution and later subjected to the tensile test. It was observed that the failure of the tensile specimen again took place in the Type 321 ASS BM, in a similar manner as in Sec. III.F. Moreover, the TS reduction was found to be in reduced amount as compared to obtained in the case without immersing in a corrosive environment (610 MPa). However, the SEM fractograph of the fractured surface showed a slight transition in the failure mechanism as shown in Fig. 7(a). The fractograph showed a large amount of ductile dimples along with cleavage facets. This showed that the continuous exposure of dissimilar welds in chloride environment was not enough to cause any sign of pitting and inter-granular corrosion even in weaker parent material that is Type 321 ASS BM. Therefore, Type 321 ASS would be a better alternative to replace Hastelloy C-276 in such chloride environment.

FIG. 7.

(a) SEM fractograph of tensile specimen tested in corrosive solution; (b) polarization curve of Hastelloy C-276 BM, Type 321 ASS BM and FZ.

The potentiodynamic polarization of both the BMs and FZ were carried out in 0.5 M H₂SO₄ + 0.01 M NH₄SCN solution at RT as depicted in Fig. 7(b). The addition of acidified ammonium thiocyanate to H₂SO₄ was found to be helpful in simulating an oxidizing medium so that the transition from an active-to-passive region of the BMs could be easily investigated. ^{Reference Pujar, Dayal, Malhotra and Gill29} It was observed that both the FZ and Hastelloy C-276 BM displayed better passivation and this could be attributed to the presence of Mo and W in the matrix. The E _pit value was lower for FZ compared to Hastelloy C-276 BM and Type 321 ASS BM. The reason for this could be attributed to the presence of Mo-rich secondary particles which might have weakened the matrix and consequently formed the basis for such pits to occur. ^{Reference Pulkkinen, Apajalahti, Papula, Talonen and Hanninen30} Moreover, Type 321 ASS BM being a comparatively inferior material had shown good resistance to the increase in current density followed by formation of the passive film. This might be due to the higher amount of Cr content in the alloy composition which contributed for such behavior in the passive film formation. Hence, the choice of replacing Hastelloy C-276 by cost-effective Type 321 ASS would be a great alternative for industrial applications.