Fusion welding of dissimilar metals constitutes a crucial processing stage in a variety of applications, and the use of high energy beams (HEB) like lasers and electron beams for such welding applications has several advantages, such as, precision, narrow heat affected zone, and consequently, low distortion. An understanding of microstructural evolution in the weld is a prerequisite for producing sound joints with desired properties. HEB welding of similar metals have been studied extensively. In contrast, fewer studies have been directed toward understanding the fundamental aspects of solidification of dissimilar welds. This thesis presents an effort in that direction by exploring microstructural evolution in Ti/Ni dissimilar welds.
Welding of Ti/Ni serves to illustrate the fundamental differences that distinguish dissimilar welding from the welding of similar metals. These are: (i) Thermophysical properties of the base metals are, in general, different, and this can have important consequences in the heat transfer conditions. (ii) Composition can vary over an wide range, the extreme being for the case of a pure binary couple, and the solid–liquid interface cannot be defined by a single liquidus isotherm. (iii) In addition to the surface energy driven Marangoni convection, a strong solutal convection can arise due to a large difference in the density of the base metals. (iv) Nucleation of phases assumes greater importance, especially in systems with intermediate phases.
We have carried out laser and electron beam welding (LW and EBW) experiments in a butt welding geometry to join Ti/Ni dissimilar couples. Weld microstructures were characterised using scanning and transmission electron microscopy (SEM and TEM); composition information was obtained from energy dispersive spectroscopy (EDS) of Xrays in the SEM. In addition to the pure binary couple, we have also studied electron beam welding of Ti/Ni with a thin Ta interlayer. We summarise our findings in each set of experiments in the following sections.
Laser welding of Ti/Ni
We have studied partial penetration welds obtained within the range of experimental parameters used in our study. These welds show the following interesting features:
1. The welds are asymmetric with respect to the initial joint. Despite its higher melting point, Ti melts more than Ni due to its lower thermal diffusivity, making the average composition of the weld richer in Ti (Ti–40at.%Ni).
2. Composition changes very steeply near the fusion interfaces in both Ti and Ni with associated microstructural changes. The variation is of much lesser magnitude in the rest of the weld, reflecting a well mixed melt pool on a macroscopic scale.
3. Growth of base metal grains into the weld pool at the fusion interfaces is severely restricted at both Ti and Ni ends.
4. The Ti fusion interface is marked by a band consisting of Ti2Ni dendrites which grow toward the Ti base metal.
5. Layered structures form at the Ni fusion interface. The sequence of the layers is: solid solution (Ni)→ Ni3Ti→ Ni3Ti+NiTi eutectic → NiTi. We note the absence of the (Ni)+Ni3Ti eutectic in this sequence.
6. NiTi and Ti2Ni are the major phases that appear in the bulk of the weld. Volume fraction and morphology of NiTi vary almost periodically to form microstructural bands.
7. Solid state transformation of NiTi results in the formation of the Rphase and martensite, which reflect the composition heterogeneity in the weld. Sometimes, Ni4Ti3 precipitates are observed also, providing indirect evidence of nonequilibrium solidification.
8. Nitrogen pickup from the atmosphere during welding leads to the formation titanium nitride dendrites in the weld.
9. Solutal convection and buoyancy forces manifest themselves through the segregation of the lighter nitride and Ti2Ni phases toward the top surface of the weld; the heavier liquid forms blocky NiTi in the bottom half of the weld.
These observations stand in striking contrast with the microstructures of conventional welds. We have proposed a set of composition and temperature profiles in the weld which reflect the diffusive and advective transport processes; when combined with thermodynamic information from the Ti–Ni phase diagram to yield spatial liquidus temperature profiles, these profiles can adequately explain most of the results. Our observations illustrate the importance of (a) nucleation, and (b) the inhomogeneous nature of the melt in which growth takes place. They also highlight the role of convective currents in bringing about local fluctuations in composition and temperature leading to ‘low velocity bands’.
Electron beam welding of Ti/Ni
We have carried out full penetration EBW of thin plates of Ti and Ni. The major observations are: (i) Average composition of the weld is in the Ni–rich side of the phase diagram (Ni–40at.%Ti). (ii) Fusion interface microstructures are very similar to that in LW exhibiting restricted base metal growth (although little amount of epitaxy can be seen in the Ni side), growth of Ti2Ni dendrites toward the base metal at the Ti fusion interface and the sequence of layers at the Ni interface: (Ni)→ Ni3Ti→ Ni3Ti+NiTi. Unlike LW, however, Ni3Ti, instead of NiTi, reappeared after the third layer on the Ni side. (iii) General microstructure consists of the Ni3Ti+NiTi eutectic, which appears in several anomalous as well as regular morphologies. (iv) Formation of NiTi is restricted mostly to regions near the Ti fusion interface. (v) Segregation of Ni3Ti was observed in a few places. The most prominent change in the microstructure compared to LW is a shift from the Ti2Ni– NiTi phases in the bulk of the weld to a Ni3Ti+NiTi eutectic structure. This is a direct consequence of the shift in the average composition of the weld to the Ni– rich side. The occurrence of different anomalous and regular eutectic structures bear similarity with bulk undercooling experiments conducted on eutectic systems having a strongly faceting phase as one of its constituents. The asymmetric coupled zone, along with composition and temperature fluctuation due to fluid flow, can be attributed to the origin of these structures.
Electron beam welding of Ti/Ni with a Ta interlayer
Motivated by the report of superior mechanical properties of Ti/Ni welds with an interlayer of Ta, whose melting point is much higher than those Ni and Ti, we performed EBW experiments using a Ni–Ta– Ti configuration. The key observations are: (i) The process is inherently unsteady in nature, and results in partial and irregular melting of the Ta interlayer. This partial melting essentially divides the weld into Ni–rich and Ti–rich halves. (ii) Microstructure near the fusion interface in Ni and Ti show similarities with that of the pure binary Ti/Ni welds; the phases here, however, contain Ta as a ternary addition. (iii) Microstructure in the Ti–rich half consists of dendrites of the Ni(Ti,Ta) phase with a high Ti:Ta ratio, and an eutectic formed between this phase and a (Ti,Ta)2Ni phase having significant amount of Ta. Two Ni(Ti,Ta) type phases dominate the microstructure in the Ni–rich half: the phase having a higher Ti:Ta ratio forms cells and dendrites, whereas the one of a lower Ti:Ta ratio creates an interdendritic network. (iv) Regions near the unmolten Ta layer in the middle show the formation of a sawtoothlike Ta–rich faceted phase of composition (Ta,Ti)3Ni2. Since very scarce thermodynamic data exist for the Ni–Ta–Ti ternary system, we have taken cues from the binary phase diagrams to understand the microstructural evolution. Such extrapolation, although successful to some extent, fails where phases which have no binary equivalents start to appear.
In summary, in this thesis, we explore microstructural evolution in the Ti/Ni dissimilar welds under the different settings of laser and electron beam welding processes. This study reveals a variety of phenomena occurring during dissimilar welding which lead to the formation of an extensive range of microstructural features. Although a few questions do remain, most results can be rationalised by drawing from, and extending the knowledge gained from previous studies by introducing physical and thermodynamic arguments.
| Identifer | oai:union.ndltd.org:IISc/oai:etd.ncsi.iisc.ernet.in:2005/601 |
| Date | 07 1900 |
| Creators | Chatterjee, Subhradeep |
| Contributors | Chattopadhyay, K, Abinandanan, T A |
| Source Sets | India Institute of Science |
| Language | en_US |
| Detected Language | English |
| Type | Thesis |
| Relation | G20950 |
Page generated in 0.0034 seconds