- Email: [email protected]
stainless steel joints
Accepted Manuscript Title: Microstructure and mechanical properties of resistance-welded NiTi/stainless steel joints Authors: Qiao Li, Yuanxiang Zhu, Jialin Guo PII: DOI: Reference:
S0924-0136(17)30273-X http://dx.doi.org/doi:10.1016/j.jmatprotec.2017.07.001 PROTEC 15298
To appear in:
Journal of Materials Processing Technology
Received date: Revised date: Accepted date:
24-1-2017 30-6-2017 1-7-2017
Please cite this article as: Li, Qiao, Zhu, Yuanxiang, Guo, Jialin, Microstructure and mechanical properties of resistance-welded NiTi/stainless steel joints.Journal of Materials Processing Technology http://dx.doi.org/10.1016/j.jmatprotec.2017.07.001 This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
Microstructure and mechanical properties of resistance-welded NiTi/stainless steel joints Qiao Li, Yuanxiang Zhu, Jialin Guo School of Power and Mechanical Engineering, Wuhan University, Wuhan 430072, China Corresponding author: Yuangxiang Zhu; email: [email protected]; tel: +86-13554026788
Abstract Resistance welding was employed to join NiTi and stainless steel (SS), with the aim to fabricate a NiTi/stainless steel joint for biomedical applications. The effects of welding current and post-welding cold drawing with 30% area reduction on the joint microstructures and mechanical properties were investigated. At a welding current of 40 A, a NiTi/SS joint with a reaction layer of several microns in size was obtained. The tensile strength of this joint reached 440 MPa, with 7.9% rupture elongation and a fracture on the NiTi side that occurred via micro-void coalescence mechanism. Good plasticity and bending performance of the joint were verified by the bending test. After cold drawing, the microstructures in the heat-affected zones (HAZs) were refined and the microhardness in the HAZs increased. The tensile strength of the joint increased to 830 MPa, with 6.2% elongation. The 40-μm-thick weld obtained at the welding current of 45 A consisted of a reaction layer and a NiTi molten zone. Local embrittlement occurred near the NiTi fusion line owing to grain coarsening, the existence of re-solidified grain boundaries in the HAZ and eutectics in the molten zone. The tensile strength of the joint was established as 340 MPa, with 5.8% rupture elongation and a cleavagemediated fracture on the NiTi side. This type of joint could not be drawn as a result of its reduced ductility. Keywords: NiTi; Stainless steel; Resistance welding; Microstructure
1. Introduction In biomedical sphere, joining of NiTi and stainless steel (SS) enables the fabrication of composite guidewire, venous catheter and orthodontic arch wire. Reliable joining techniques that can prepare NiTi/SS joints with good mechanical properties, especially bending performance, are thus of great importance. The joining methods reported currently, however, are not satisfactory. The brittle nature of the NiTi/stainless steel (SS) joint was described for the first time by Eijk et al. (2003) in a preliminary study on plasma welding of NiTi and SS. Li et al. (2005) employed capacitor discharge upset welding to join NiTi and SS. The authors reported that while both of the base metals melted, the liquid phase was not removed from the NiTi/SS interface completely, thus resulting in low strength (150 MPa). Fukumoto et al. (2010) used friction welding to prepare NiTi/SS joints, demonstrating that welding without Ni interlayer results in the formation of brittle Fe2Ti phase at the weld interface. This outcome suggested that the solidstate bonding over several seconds failed to inhibit the formation of brittle phases. Gugel et al. (2008) reported the mechanical behaviour of laser-welded NiTi/SS joints, with strengths of up to 623 MPa, but failed to provide any microstructural information.
In order to inhibit the formation of intermetallic phases, Li et al. (2011) introduced Ag–Cu foil as a filler metal in transient liquid phase diffusion bonding (TLPDB). Nevertheless, undesirable phases continued to form in the interfacial regions, rendering the results less than satisfactory. In a comprehensive series of studies on laser welding of NiTi and SS by Li and co-workers, Ni (Li et al. (2012)), Co (Li et al. (2013a)), and Cu (Li et al. (2013b)) were introduced as filler metals. Although the formation of Fe–Ti and Cr–Ti intermetallic phases was inhibited effectively, the excessive addition of filler metals was found to deteriorate the joint properties by inducing proliferation of other brittle phases. The maximum joint strength obtained in the work by Li and co-workers was established as 520 MPa, with 5.1% rupture elongation. In a dissimilar metal weld joint, the existence of brittle intermetallic phases does not necessarily lead to the deterioration of joint properties. For example, in diffusion bonding of Ti to 304 SS, Ghosh et al. (2003) showed that when the growth of the intermetallic compound layer, which consisted mainly of FeTi and Fe2Ti phases, was suppressed to 7 m through temperature optimisation, the tensile strength reached 70% of that of the titanium base metal, with 9% rupture elongation, which indicated good ductility. In dissimilar welding of aluminium alloy to 304 SS, Murakami et al. (2003) showed that when the thickness of the intermetallic layer was decreased to nanometre size, the joint strength exceeded that of the Al base metal. In resistance welding, the formation of brittle phases can be easily suppressed through parameter optimisation, and the detrimental effects of brittle phases can thus be minimised. In order to obtain NiTi/SS joints with good mechanical properties, the following measures were taken in this work:
1) The detrimental effects of intermetallic phases were minimised through adjustment of the welding current; 2) A SS tube was used to improve the strength and bending performance; 3) Post-welding cold drawing was employed to recover softened heat-affected zones (HAZs). In this study, the chemical compositions, microstructures, and microhardness values of the prepared NiTi/SS joints are examined and discussed, and the effects of welding current and cold drawing on microstructures and mechanical properties are investigated.
2. Experimental procedures 2.1 Materials and sample preparation NiTi (50.9 at.% Ni) and AISI304 SS were purchased from Fort Wayne Metals. The wire and tube sizes are marked in Fig. 1. The as-received NiTi wires were subjected to 45 ± 5% cold drawing, followed by superelastic heat treatment. The as-received SS wires were subjected to 93% cold drawing, followed by stressrelieving annealing. All wires were sanded to a bevelled tip. Oxide coatings were removed by electropolishing in a mixed solution of perchloric acid and acetic acid (in a volume ratio of 1:10) at a voltage of 22 V. The wires were ultrasonically cleaned in an acetone bath for 5 min and dried in atmosphere prior to welding.
Fig. 1. A snapshot (left) of the joint assembly and corresponding schematic cross-section (right). The red line in the schematic representation denotes the electrical route. The two bevelled-tip wires are inserted symmetrically into an SS tube, which is surrounded with compacted Al2O3 powder for cooling and constraining purposes. Powder compacting is performed using a custom-built device. Copper electrodes are installed at the end of the SS tube, yielding a sandwich-like assembly. The sandwich-like assembly is mounted onto a custom-made welding mould. The wires in this mould are clamped in order to permit the application of upset pressure.
2.2 Resistance welding As shown in Fig. 1, NiTi and SS wires were inserted into an SS tube at an equal offset, in a manner that enabled their bevelled tips to come into contact with each other. The tube-wire assembly was mounted onto a custom-made device, with which Al2O3 powder was added around the SS tube and compacted under a force of around 1000 N. A copper electrode was installed at both ends of the SS tube after the powder compacting. The sandwich-like assembly was subsequently mounted onto a custom-designed welding mould. The wires in this mould were fixed with clamps and the copper electrodes were connected to a power supply. Upset pressure was applied through the clamps (maximum value of 500 MPa). Resistance welding was achieved by passing a single electrical pulse along the route illustrated in Fig. 1 (red line). The welding experiments were divided into three groups according to the current employed, i.e. 37 A, 40 A and 45 A, and were conducted in a vacuum chamber. Current duration was set to 100 ms and the upset pressure to 270 MPa, i.e. values above which the wire in the unconstrained region (the region between the wire clamp and the copper electrode) would deflect. The 40 A joints (i.e. joints produced at a current of 40 A) were cold drawn from 0.43 mm to 0.36 mm after welding, resulting in a 30% area reduction. For the 45 A joint, however, the drawing process typically resulted in failure because the material was too brittle. For the 37 A joint, only several welding spots were observed at the interface. As a result of a low interfacial strength, NiTi was frequently pulled out from the tube. Microstructural and mechanical information about this type of joint is provided in the supplementary information.
2.3 Metallography and characterisation To analyse the microstructures of the polished joint specimens, NiTi was etched using a mixture of 10% HF, 40% HNO3 and 50% H2O, and SS was etched in a mixture reagent of HCl and HNO3 in a ratio of 3:1. Optical microstructures were visualised using a Zeiss LAB1 microscope. To analyse the weld microstructures, the polished samples were examined by scanning electron microscopy (SEM, JEM2012-FEF) in the backscattered mode. The chemical composition analysis was performed using energy-dispersive spectroscopy (EDS). Microhardness measurements were performed using a Vickers hardness tester with a load of 200 g and a dwell time of 10 s. Fig. 2 shows a schematic representation of the performed hardness distribution measurements. The mean hardness of the base metals was calculated from values measured at five different locations. Static tensile tests were performed using a MTS810 material testing system. In these experiments, the cross-head displacement and the sample gauge length were set to 0.3 mm/min and 50 mm, respectively. The tensile stress in the strain-stress curves was calculated using the external diameter of the SS tube, i.e. 0.43 mm for the as-welded joints and 0.36 mm for the as-drawn joints. The tensile strength was determined as the average of five joints per condition. Finally, SEM was used to examine the fracture morphology.
Fig. 2. Schematic representation of the hardness distribution measurement for each joint. 3. Results and discussion 3.1 Chemical compositions and SEM microstructures of the as-welded joints Fig. 3 shows the SEM micrographs of the 40 A joint obtained in the backscattered mode. It can be observed from Fig. 3b that with the exception of the sharp corners, metallurgical bonding formed at the bevelled NiTi/SS interface. Under upset pressure, the NiTi material fused with the SS tube, as shown in Fig. 3e. As a result of the temperature gradient induced by copper electrodes, the weld shown in Fig. 3c varied in thickness (~1–10 μm) along the bevelled interface. The highest temperature was close to the central point of the bevelled interface, thus making the weld closest to this location the thickest. As shown in Fig. 3c–e, the NiTi/SS weld consisted of dark-grey grains, which exhibited predominantly worm-like morphology, and a light-grey matrix. A subtle gap (~1 μm in thickness) was observed in zone C, as shown in Fig. 3f, indicating insufficient diffusion. The results of EDS analysis of the 40 A joint are shown in Fig. 4. It is apparent from Fig. 4a that in zone A, the content distribution of Fe displays the same trend as the distribution of Cr, and the inverse trend relative to Ni. Ti distribution did not exhibit any apparent correlation with the content distributions of other
components. In spite of the content fluctuation, the peaks in the element composition profiles remained mostly stable, and similar behaviour was observed for the valleys. This alternating pattern suggested that the dark grains and the light-grey matrix arise from two different phases. The element composition profiles in zone C changed smoothly and rapidly across the weld, with no major fluctuations observed in Fig. 4b, which indicates that atomic diffusion has taken place.
Fig. 3. Backscattered SEM micrographs of the 40 A joint. Contrast enhancing was applied to images in (c–d) using Phtotoshop. The image in (a) shows the outline of the joint while (b) focuses on the weld, with magnified views of zones A, B and C provided in (c), (e) and (f), respectively. The magnified view in (d) shows the microstructure of the weld. EDS point analysis was carried
out on the worm-like phase (point 1 in (d)) and inter-grain light-grey matrix (point 2 in (d)).
EDS point analysis was carried out on points marked 1–4 in Fig. 3. Chemical compositions were stamped on the 1000 ºC isothermal section of the Fe–Ni–Ti ternary system, as reported by Cacciamani et al. (2006) and illustrated in Fig. 4c. Quantitative determination of temperature in the current work was difficult because of the very small size of specimens and the closed nature of the welding mould. In solid-state resistance welding, the temperature usually needs to reach ~80–90% of the melting point of the base metal in order to achieve metallurgical bonding. The isothermal section at 1000 ºC was therefore selected in this work. The high cooling rate, however, did not allow sufficient time for atomic diffusion during cooling. As a consequence, the phases formed during the heating stage were likely to be retained at room temperature, regardless of the influence of cooling. Among the main elements detected in the two base metals, the content of Cr was the lowest. For this reason, Cr was expected to have minimal influence on the chemical composition and phase formation of the weld. Therefore, only Ti, Fe and Ni were considered in phase deduction. According to the diffusion path theory suggested by Kirkaldy and Brown (1963), the phases that can be produced in welds are those present in the phase domains crossed by the line connecting the two base metals (Fig. 4c, blue line). As shown in Fig. 4c, therefore, the potential phases produced in the NiTi/SS weld in the present scenario are (Fe, Ni) Ti, Ni3Ti, Fe2Ti and γ-(Fe, Ni). According to the Cr–Ti–Ni ternary phase diagrams reported by Tan et al. (2007) and Cr–Ti–Fe ternary phase diagrams provided by Wang et al. (2017), the solubility of Cr is high in Fe2Ti (>60 at.%) but low in Ni3Ti (11.4 at.%) and NiTi (14.5 at.%). Taking these facts into consideration, the Cr in the NiTi/SS weld was thought to dissolve in Fe2Ti, and the Cr content was linked to the Fe content. This hypothesis was supported by the fact that the Fe and Cr content profiles displayed the same trends in behaviour (Fig. 4a).
Fig. 4. EDS line-scan results obtained for zone A (a) and zone C (b). (c) EDS analysis of the points marked in Fig. 3, stamped on the 1000 ºC isothermal section of a Fe–Ni–Ti phase diagram according to a procedure described by Cacciamani et al. (2006). The blue line in the isothermal section connects the two base metals, and indicates the phases that are potentially present in NiTi/SS joints. According to the diffusion path theory reported by Kirkaldy and Brown (1963), the possible phases are those present in the phase domains crossed by the blue line connecting the two base metals. In addition to the main phases from the base metals, the intermetallic phases that can be present in the joint are thus Fe2Ti and Ni3Ti.
It is clear from Fig. 4c that point 1, which corresponds to the worm-like phase shown in Fig. 3d, was most likely comprised of Fe2Ti, while point 2, corresponding to the light grey matrix in Fig. 3d, was probably a eutectic comprised of (Fe, Ni)Ti/Ni3Ti/Fe2Ti. Points 3 and 4 in zone C corresponded most likely to γ-(Fe, Ni) and (Fe, Ni)Ti, respectively. The worm-like morphology suggested that the weld experienced liquation. The melting point of the weld was probably lowered by the complexity of its chemical composition. According to
the reaction scheme described by Cacciamani et al. (2006) for the Fe–Ni–Ti system, worm-like Fe2Ti was probably produced as the proeutectic phase in the (Fe, Ni)Ti/Ni3Ti/Fe2Ti eutectic reaction. As a result of the growth of worm-like Fe2Ti, the chemical composition of the inter-grain region reached the eutectic point gradually, thereby producing the ternary eutectic. Backscattered SEM micrographs of the 45 A joint are shown in Fig. 5. The bevelled interface appeared steeper than that shown in Fig. 3 for the 40 A joint, most likely as a result of the differences in the cutting angle during sanding. The weld widened to around 40 μm, as shown in Fig. 5c. Under upset pressure, a fraction of mashed NiTi flowed away, thereby leaving behind a hole, as indicated by the yellow ellipse in Fig. 5b. The SS wire-tube reaction layer (RL) in Fig. 5f was found to be rich in Ti, as evidenced by EDS analysis, which indicates that the mashed NiTi was squeezed into the wire-tube gap and reacted with the SS. Fig. 5c shows that the bevelled weld varied slightly in thickness. A magnified view provided in Fig. 5d shows that the weld consisted of two layers, i.e. a RL adjacent to the SS and a NiTi molten zone (MZ). The microstructure of the RL was similar to that observed for the 40 A weld. The dark-grey grains in RL exhibited both equiaxed and needle-like morphologies. Coarse needle-like grains dominated the RL centre, whereas fine equiaxed grains dominated the RL periphery. The NiTi MZ displayed typical melted and re-solidified microstructures, consisting of equiaxed grains and a light-grey matrix. As shown in Fig. 5e, the weld between NiTi and the SS tube also displayed a two-layer structure. The grains in MZ, however, were columnar, an observation that can be attributed to the steep temperature gradient imposed by the heat sink. Grain boundary (GB) liquation was observed in NiTi HAZ (Fig. 5e). An arrested micro-crack, caused most likely by solidification cracking along the columnar grain boundary, was observed near the liquated GBs.
Fig. 5. Backscattered SEM micrographs of the 45 A joint. Contrast enhancing was applied to images in (c–f) using Photoshop. Image in (b) shows a full view of the weld, with zones A, B and C shown at higher magnification in (c), (e) and (f), respectively. The hole highlighted in (b) was produced by mashed NiTi flowing away to the side of the SS, as indicated by the yellow arrow. The magnified view in (d) shows the two-layered structure of the weld, i.e. a reaction layer (RL) adjacent to SS and a NiTi molten zone (MZ). EDS analysis was carried out on the RL (point 1-7), the RL-MZ interfacial region (point 9-8) and the MZ (point 10-15). Grain boundary (GB) liquation was captured in (e).
Fig. 6. Element composition profiles (a) and chemical compositions stamped on phase diagram (b) across the weld. The phase diagram represents the 1000 °C isothermal section of the Fe–Ni– Ti ternary system, as described by Cacciamani et al. (2006). Point 1-7 represents the RL; point 8-9 represents the RL-MZ interfacial region; and point 10-15 represents the MZ
Element composition profiles of the 45 A joint are shown in Fig. 6a. In this system, the Ti distribution displayed major fluctuations over the entire MZ. Further, the distributions of Ni and Fe fluctuated significantly in the interfacial region between RL and MZ. The Cr distribution, by contrast, did not exhibit any significant fluctuations. As shown in Fig. 6b, the chemical compositions of points 1–15 marked in Fig. 5d were stamped on the 1000 °C isothermal section of the Fe–Ni–Ti ternary system. Points 1–7 were in the RL, points 8–9 in the RL–MZ interfacial region, and points 10–15 in the MZ. It is apparent from Fig. 6b that the chemical composition of the NiTi MZ was basically comprised of a Ni3Ti/(Fe, Ni)Ti (a solid solution of NiTi B2 phase, in which Fe atoms substitute a portion of Ni atoms) two-phase domain. Although it is possible that Fe2Ti might exist in the MZ, its presence would represent only a minor phase since the chemical composition of MZ was considerably different from that of the Fe2Ti single-phase domain. Taking into consideration this comparison of chemical compositions, a mechanism for the phase formation in the MZ can be therefore proposed: when molten NiTi starts to solidify, the (Fe, Ni)Ti phase, corresponding to the dark-grey grains in the MZ, grows epitaxially on the grains of the base metal. Accompanying the solidification, the liquid phase was expelled to the inter-grain region, and its chemical composition gradually reached the eutectic point, ultimately giving rise to (Fe, Ni)Ti/Ni3Ti eutectic in the inter-grain region. In the RL, points 1 and 2, corresponding to the light-grey matrix, represent most likely the γ-(Fe, Ni)/Fe2Ti eutectic. Chemical compositions of points 3 and 4 (dark-grey grains in the RL) were comprised of the γ-(Fe, Ni)/Fe2Ti/Ni3Ti three-phase domain, with Ni3Ti and γ-(Fe, Ni) as minor phases. According to the chemical compositions of
points 3–7, the dark-grey grains in the RL were probably composed of Fe2Ti. In the RL–ML interfacial region, the light-grey matrix, corresponding to points 8 and 9, was probably comprised of the Fe2Ti/(Fe, Ni)Ti/Ni3Ti eutectic.
3.2 Metallographic microstructures Fig. 7 shows the optical microstructures of the as-welded 40 A joint. Fig. 7a shows that a gap was produced at the interface as a result of preferential etching. In terms of the SS, although equiaxed grains (~10–20 μm) were prominent near the weld (Fig. 7b), elongated fibre-like microstructures were observed in the base metal (Fig. 7d). The grains shown in Fig. 7c were found to be finer than those in Fig. 7b owing to the lower heating temperature. A heating boundary can be identified in Fig. 7d. In this case, the nearby grains were so fine that it was not possible to identify them using optical microscopy. The as-received SS wires were subjected to 93% cold drawing and subsequent stress-relieving annealing at 350–427 °C. The existence of elongated microstructure in the SS base metal indicates that the stress relieving annealing changed the subgrain structures through dislocation annihilation and polygonisation. According to Naghizadeh and Mirzadeh (2016), the microstructures of the SS base metal consist mainly of stress-induced martensite, and will experience the following processes during heating: 1) reversion of stress-induced martensite to austenite, 2) recrystallisation of retained austenite and 3) growth of the reversed and recrystallised austenite grains. It can be seen from Fig. 7d that the elongated microstructures on the right-hand-side of the heating boundary disappeared, indicating that the reversion process has started. Closer to the weld (Fig. 7c), fine equiaxed grains (~4 m) were observed, which suggests that the reversion and recrystallisation processes have finished. Near the weld, the reversed and the recrystallised austenite grains grew to a length of ~10–20 m, as shown in Fig. 7b.
Fig. 7. Optical microstructures of the as-welded 40 A joint: full view (a). The zones marked in (a) are shown in more detail in the correspondingly labelled images. Images shown in (d) and (i) were captured near the end of the SS tube (these locations could not be included in the image shown in (a)).
In terms of NiTi, the grains were found to be ~10 m in size near the weld, as shown in Fig. 7e. Finer grains were also observed and these are portrayed in Fig. 7g–h. The NiTi base metal did not display any grains when examined by optical microscopy (Fig. 7i). According to the microstructural report provided for NiTi by the manufacturer (Schaffer (2009)), the as-received NiTi wires, which were subjected to 45 5% cold drawing and super-elastic annealing, consisted of nano-sized recrystallised grains (~5–60 nm). The report also stated that when the nano-grains increase in size to 10 μm, the NiTi is expected to lose its superelasticity. During welding, the grains in the NiTi HAZ increased in size, resulting in functional degradation and decline in strength. The presence of lath microstructure near the weld (Fig. 7e–f) indicates the existence of martensite, which was probably caused by the precipitation of Ni-rich phases. In this system, martensite coexisted with the B2 phase near the weld.
Fig. 8. Optical microstructures of the 45 A joint: full view (a), the bevelled weld (b–c), reaction layer (RL) between SS wire and SS tube (d), the weld between NiTi and SS tube (e), NiTi HAZ (f) and SS HAZ (g). The RL and the molten zone (MZ) in images (b) and (e) correspond to those shown in Fig. 5d. The image in (c) shows the solidification mode transition in NiTi MZ. Image in (i) reveals grain boundary (GB) liquation in the NiTi HAZ.
The optical microstructures of the 45 A joint are shown in Fig. 8. As shown in Fig. 8b, it was possible to identify both the RL and MZ in the etched weld. Several holes were observed in the RL as a result of preferential etching, as shown in Fig. 8a and Fig. 8c. The dark-etched RL eutectic shown in Fig. 8b was in good agreement with the results of EDS analysis discussed in the previous section. Fig. 8c shows that the coarseness of the grains increased to ~20 m near the NiTi fusion line. The solidified grains in the MZ grew epitaxially on the base metal, indicating that they had the same crystal structure as the NiTi base metal itself. This observation further corroborates the phase formation mechanism proposed in the previous section. Analyses of the chemical compositions revealed that the solidified equiaxed grains were comprised of (Fe, Ni)Ti (i.e. a solid solution of B2 phase). The (Fe, Ni)Ti grains grew first epitaxially from the fusion line, and in the planar mode along the easy-growth direction of the base metal grains. A short distance away from the fusion line, however, solidification changed to the cellular mode. Deeper into the weld centre, equiaxed dendrites nucleated and grew. This change in solidification mode was caused by increasing constitutional
supercooling in the direction from the fusion line to the weld centre. Fig. 8e shows that the MZ was almost twice as thick as the RL, and the solidification mode changed from planar to cellular, columnar dendritic and equiaxed dendritic. Fig. 8f displays GB liquation that was observed within ~100 μm of the weld, which was caused by the direct melting of the matrix since there was no pre-existing eutectic with a low melting point in the NiTi. According to Kou (2003), the large cooling rate during the welding process can delay the solidification of melted GBs until the eutectic temperature is reached, thereby leading to the formation of an inter-grain eutectic. In the present case, (Fe, Ni)Ti/Ni3Ti eutectic was probably produced after GB solidification. In the SS HAZ (Fig. 8g–h), the grains did not exhibit a further increase in coarseness when compared to those observed in the as-welded 40 A joint. This observation indicates that even though the temperature increased, the welding time was too short for further coarsening. The full RL shown in Fig. 8d reveals a dark-etched eutectic.
Fig. 9. Deformed SS (a) and NiTi (c) at either end of the SS tube, and microstructures (b, d–f) of the as-drawn 40 A joint.
Fig. 9 shows the optical microstructures and deformed wires of the as-drawn 40 A joint. The 40 A joints were cold drawn from 0.43 mm to 0.36 mm, which corresponds to a 30% area reduction. The slope of the bevelled NiTi/SS interface decreased after cold drawing, as shown in Fig. 9b. Fig. 9a shows that no evident microstructural changes were observed at the onset of SS deformation. By contrast, the deformation onset resulted in the formation of a dark boundary on the NiTi side (Fig. 9c). A magnified image shown in Fig. 9f reveals that the dark boundary consisted of a needle-like phase. At this stage, however, it was unclear as to what the needle-like phase was comprised of and why it formed. Fig. 9d shows that the austenite grains in SS HAZ decreased in size to ~5 μm as a consequence of cold drawing and the grain boundaries became more obscure. In terms of NiTi, the grains in the HAZ decreased in size to less than 1μm and became unidentifiable by optical microscopy, as shown in Fig. 9e. In fact, only a scattered needle-like phase was observed
throughout the NiTi HAZ. Microstructure refinement suggested that although the cold drawing resulted in partial recovery of the softened HAZ, further plastic deformation would be required for complete recovery.
3.3 Microhardness and mechanical properties Fig. 10 shows the microhardness distributions and Vickers indentations of NiTi/SS joints. The values of mean hardness for the NiTi and SS base metals were determined as 405 HV0.2 and 578 HV0.2, respectively. For the as-welded 40 A joint, the mean hardness throughout the entire NiTi HAZ was 302 HV0.2. Hardness in the SS HAZ remained stable (189 HV0.2 on average) near the weld, and increased subsequently (in direction away from the weld) to the value of the base metal. Minimum hardness of the joint was established as 172 HV0.2, and existed on the SS side. The hardness profile of the as-welded 40 A joint did not exhibit the pattern characteristic for interlayer-free NiTi/SS joints obtained in previous investigations by Li et al. (2013a), Li et al. (2013b), Li et al. (2012) and Ng et al. (2015), in which a high-hardness plateau was observed in the hardness profile of the welds. As shown in Fig. 10b, the hardness at the NiTi/SS interface was somewhere in between those of the two base metals. After cold drawing, a dramatic increase was observed in the hardness of the as-drawn 40 A joint (Fig. 10a). In fact, the NiTi HAZ was found to be almost as hard as the base metal, and the mean hardness of the SS HAZ increased to 377 HV0.2, which represents a two-fold increase relative to the as-welded joint. The increase in hardness observed for the as-drawn joint was fairly consistent with the results of microstructure refinement.
Fig.10. Microhardness profiles of NiTi/SS joints (a) and microstructures showing Vickers indentations (b–e). Image (b) shows the weld of the as-welded 40 A joint. Images in (c–e) show the bevelled weld, reaction layer between SS wire and SS tube and the weld between NiTi and the SS tube of the 45 A joint. All indentations were created using a 200 g load at a dwell time of 10 s. The hardness values associated with the different indentations are marked in each image.
For the 45 A joint, the hardness profile of the SS HAZ was similar to that of the as-welded 40 A joint. The weld hardness of the 45 A joint, however, increased relative to the 40 A joint as a result of the increased proportions of Fe2Ti and the eutectic. Within ~100 m from the weld, the hardness of the NiTi HAZ exceeded
that of the base metal, which was attributed to GB liquation. As can be seen in Fig. 10c, the highest hardness of the joint was on the NiTi fusion line and the value exceeded 800 HV0.2, thus indicating local embrittlement. The microstructure near the NiTi fusion line was exposed to the coarsest grains throughout the entire HAZ, re-solidified GBs and the eutectic in the MZ. Nevertheless, the deformation coordination between grains was impeded by re-solidified GBs and the eutectic in the MZ, thereby resulting in the embrittlement. The hardness of the RL located between the SS wire and the SS tube (Fig. 10d) indicates that the complex eutectic structures did not embrittle the material severely. The hardness on the NiTi fusion line (Fig. 10e) was determined to be only 455 HV0.2, i.e. much lower than that shown in Fig. 10c owing to the presence of finer grains (~10 μm). Owing to the columnar structure, the hardness of the NiTi-tube weld was higher than that of the bevelled weld shown in Fig. 10c.
Fig. 11. Tensile deformation behaviour of dissimilar NiTi/SS joints. Three curves are presented for each condition.
Fig. 11 demonstrates the tensile deformation behaviour of NiTi/SS joints. Average tensile strength and rupture elongation of the as-welded 40 A joint were determined to be 440 MPa and 7.9%, respectively. It is clear from the strain-stress curves that the joints experienced stress plateaus prior to their ultimate fracture. For the 45 A joint, the embrittlement near the NiTi fusion line resulted in the deterioration of mechanical properties, with the mean tensile strength and elongation declining to 340 MPa and 5.8%, respectively. For this joint, the fracture occurred before reaching the maximum strain of the stress plateau. The mean tensile strength of the as-drawn 40 A joints reached 830 MPa, with a slight decrease in the mean fracture elongation to 6.2%. The strain-stress curves of these joints show that major stress fallings occurred just before the stress plateaus, with several minor ones appearing also after the plateaus. This odd deformation behaviour is quite common for resistance-welded joints, and is most likely related to the interactions formed between the peripheral SS tube and the core wires under tensile load. The fracture morphologies of NiTi/SS joints are shown in Fig. 12. The as-welded 40 A joint was found to fracture at the root of the NiTi slope and additionally, the NiTi wire was pulled out from the SS tube (Fig.12a). Fine dimples were observed on the fracture surface shown in Fig. 12b, indicating that the joint was
broken as a result of micro-void coalescence. The side surface of the NiTi wire shown in Fig. 12c exhibited a rough morphology, with many longitudinal ridges. Under tensile load, the ridges acted as microtube-wire bonding devices, thereby enhancing the joint strength. As shown in Fig. 13, the as-welded 40 A joint was bent around a rod with a diameter of 2 mm. Instead of fracturing, however, a plastic deformation was observed after bending, thereby indicating good plasticity and bending performance.
Fig. 12. Fracture morphologies of NiTi/SS joints: the as-welded 40 A joint (a–c), the as-drawn 40 A joint (d–f) and the as-welded 45 A joint (g–i). The focus zones marked in (a), (d) and (g) are shown at higher magnification in the images labelled with corresponding letters.
Fracture morphology of the as-drawn 40 A joint is shown in Fig. 12d–f. The joint fractured at the NiTi slope root as well. Dimples dominated the fracture surface, thus indicating the occurrence of a micro-void coalescence failure. According to the microstructures and microhardness of the as-drawn joint, the cold drawing reinforced the softened HAZs. As a result of the shrinking after cold drawing, the external SS tube was able to secure the core wire more tightly. Overall, these two points improved the tensile strength of the as-drawn joint. As shown in 12g, the 45 A joint cracked on the NiTi side, a process that was accompanied by tube fracture, which could be attributed to the brittle character of the weld between the tube and NiTi. As shown in Fig. 10e, columnar grains grew perpendicularly to the tube wall, resulting in strong anisotropic mechanical
properties. Under tensile load, cracks were typically initiated in the inter-grain region where the eutectic existed, and propagated along the columnar grain boundary, extending to the tube wall and eventually resulting in its breakage. Both cleavage terraces and river patterns were observed on the fracture surfaces, which are indicative of cleavage fracture mechanism. Local embrittlement near the NiTi fusion line in particular, was the primary reason behind the cleavage fracture. The direction of the river patterns suggested that the failure initiated from the crack shown in Fig. 12h.
1 mm
Fig. 13 A bent as-welded 40 A joint. Although the bending test caused a considerable plastic deformation, the joint did not break.
4. Conclusions Herein, resistance welding was employed in order to fabricate NiTi/SS joints for biomedical applications. Effects of welding current and post-welding cold drawing on microstructures and mechanical properties of the joints were investigated. The following conclusions can be drawn from the current study:
1.
The increase in welding current widened the weld and altered its microstructure. At a welding current of 40 A, the weld was comprised of a RL with 10 m in size. At 45 A welding current, the weld widened to 40 m, with a NiTi MZ produced in addition to the RL.
2.
The weld formed between the NiTi wire and the SS tube increased the bonding area. At a welding current of 45 A, the NiTi MZ of the NiTi/tube weld displayed a columnar microstructure with some micro-cracks.
3.
For the 40 A joint, the 30% area reduction induced by the cold drawing caused a decrease in the size of the grains in the SS HAZ to 5 m, and a two-fold increase in microhardness. By contrast, the grains in the NiTi HAZ decreased in size to less than 1 m and the microhardness remained similar to that of the base metal.
4.
The tensile strength and rupture elongation of the 40 A joint were established as 440 MPa and 7.9%, respectively. After cold drawing, the tensile strength increased to 830 MPa, while the
rupture elongation decreased slightly to 6.2%. The tensile strength and rupture elongation of the 45 A joint were determined as 340 MPa and 5.8%, respectively. Acknowledgement Funding: This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.
References Cacciamani, G., De Keyzer, J., Ferro, R., Klotz, U.E., Lacaze, J., Wollants, P., 2006. Critical evaluation of the Fe–Ni, Fe– Ti and Fe–Ni–Ti alloy systems. Intermetallics 14, 1312-1325. Eijk, C.V.D., Fostervoll, H., Sallom, Z.K., Akselsen, O.M., 2003. Plasma welding of NiTi to NiTi, stainless steel and hastelloy C276. Joining of Advanced and Specialty Materials VI, 125-129. Fukumoto, S., Inoue, T., Mizuno, S., Okita, K., Tomita, T., Yamamoto, A., 2010. Friction welding of TiNi alloy to stainless steel using Ni interlayer. Sci Technol Weld Joi 15, 124-130. Ghosh, M., Bhanumurthy, K., Kale, G.B., Krishnan, J., Chatterjee, S., 2003. Diffusion bonding of titanium to 304 stainless steel. Journal of Nuclear Materials 322, 235-241. Gugel, H., Schuermann, a., Theisen, W., 2008. Laser welding of NiTi wires. Materials Science and Engineering: A 481482, 668-671. Kirkaldy, J.S., Brown, L.C., 1963. Diffusion Behaviour in Ternary, Multiphase Systems. Canadian Metallurgical Quarterly 2, 2:89. Kou, S., 2003. Welding Metallurgy, 2nd Edition. 310-346. Li, H., Li, Z.X., Wang, Y.L., Feng, J.C., 2011. Transient Liquid Phase Diffusion Bonding of TiNi Shape Memory Alloy and Stainless Steel. Rare Metal Mat Eng 40, 1382-1386. Li, H., Sun, D., Cai, X., Dong, P., Gu, X., 2013a. Laser welding of TiNi shape memory alloy and stainless steel using Co filler metal. Optics & Laser Technology 45, 453-460. Li, H., Sun, D., Gu, X., Dong, P., Lv, Z., 2013b. Effects of the thickness of Cu filler metal on the microstructure and properties of laser-welded TiNi alloy and stainless steel joint. Materials & Design 50, 342-350. Li, H.M., Sun, D.Q., Cai, X.L., Dong, P., Wang, W.Q., 2012. Laser welding of TiNi shape memory alloy and stainless steel using Ni interlayer. Materials & Design 39, 285-293.
Li, M., Sun, D., Qiu, X., Sun, D., Yin, S., 2005. Microstructures and properties of capacitor discharge welded joint of TiNi shape memory alloy and stainless steel. China Welding, 95-100. Murakami, T., Nakata, K., Tong, H., Ushio, M., 2003. Dissimilar Metal Joining of Aluminum to Steel by MIG Arc Brazing Using Flux Cored Wire. Transactions of the Iron and Steel Institute of Japan 43, 1596-1602. Naghizadeh, M., Mirzadeh, H., 2016. Microstructural Evolutions During Annealing of Plastically Deformed AISI 304 Austenitic Stainless Steel: Martensite Reversion, Grain Refinement, Recrystallization, and Grain Growth. Metall Mater Trans A 47A, 4210-4216. Ng, C.H., Mok, E.S.H., Man, H.C., 2015. Effect of Ta interlayer on laser welding of NiTi to AISI 316L stainless steel. Journal of Materials Processing Technology 226, 69-77. Schaffer, J.E., 2009. Structure-Property Relationships in Conventional and Nanocrystalline NiTi Intermetallic Alloy Wire. J Mater Eng Perform 18, 582-587. Tan, Y.-h., Xu, H.-h., Du, Y., 2007. Isothermal section at 927 °C of Cr-Ni-Ti system. Transactions of Nonferrous Metals Society of China 17, 711-714. Wang, S., Wang, K., Chen, G., Li, Z., Qin, Z., Lu, X., Li, C., 2017. Thermodynamic modeling of Ti-Fe-Cr ternary system. Calphad 56, 160-168.
Figure and Table Captions. 1.
A snapshot (left) of the joint assembly and corresponding schematic cross-section (right).
2.
Schematic of hardness distribution measurement for the joint.
3.
Back scattered micrographs of the 40 A joint.
4.
EDS analysis results of the 40 A joint.
5.
Back scattered micrographs of the 45 A joint.
6.
EDS analysis results of the 45 A joint.
7.
Optical microstructures of the 40 A as-welded joint.
8.
Optical microstructures of the 45 A joint.
9.
Optical microstructures of the 40 A as-drawn joint.
10. Hardness profiles and Vickers indentations of NiTi/SS joints. 11. Tensile performance of NiTi/SS joints.
12. Fracture morphology of NiTi/SS joints. 13. A bent 40 A as-welded joint.
Figr-1
Figr-2
Figr-3
Figr-4
Figr-5
Figr-6
Figr-7
Figr-8
Figr-9
Figr-10
Figr-11
Figr-12
Figr-13
S0924-0136(17)30273-X http://dx.doi.org/doi:10.1016/j.jmatprotec.2017.07.001 PROTEC 15298
To appear in:
Journal of Materials Processing Technology
Received date: Revised date: Accepted date:
24-1-2017 30-6-2017 1-7-2017
Please cite this article as: Li, Qiao, Zhu, Yuanxiang, Guo, Jialin, Microstructure and mechanical properties of resistance-welded NiTi/stainless steel joints.Journal of Materials Processing Technology http://dx.doi.org/10.1016/j.jmatprotec.2017.07.001 This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
Microstructure and mechanical properties of resistance-welded NiTi/stainless steel joints Qiao Li, Yuanxiang Zhu, Jialin Guo School of Power and Mechanical Engineering, Wuhan University, Wuhan 430072, China Corresponding author: Yuangxiang Zhu; email: [email protected]; tel: +86-13554026788
Abstract Resistance welding was employed to join NiTi and stainless steel (SS), with the aim to fabricate a NiTi/stainless steel joint for biomedical applications. The effects of welding current and post-welding cold drawing with 30% area reduction on the joint microstructures and mechanical properties were investigated. At a welding current of 40 A, a NiTi/SS joint with a reaction layer of several microns in size was obtained. The tensile strength of this joint reached 440 MPa, with 7.9% rupture elongation and a fracture on the NiTi side that occurred via micro-void coalescence mechanism. Good plasticity and bending performance of the joint were verified by the bending test. After cold drawing, the microstructures in the heat-affected zones (HAZs) were refined and the microhardness in the HAZs increased. The tensile strength of the joint increased to 830 MPa, with 6.2% elongation. The 40-μm-thick weld obtained at the welding current of 45 A consisted of a reaction layer and a NiTi molten zone. Local embrittlement occurred near the NiTi fusion line owing to grain coarsening, the existence of re-solidified grain boundaries in the HAZ and eutectics in the molten zone. The tensile strength of the joint was established as 340 MPa, with 5.8% rupture elongation and a cleavagemediated fracture on the NiTi side. This type of joint could not be drawn as a result of its reduced ductility. Keywords: NiTi; Stainless steel; Resistance welding; Microstructure
1. Introduction In biomedical sphere, joining of NiTi and stainless steel (SS) enables the fabrication of composite guidewire, venous catheter and orthodontic arch wire. Reliable joining techniques that can prepare NiTi/SS joints with good mechanical properties, especially bending performance, are thus of great importance. The joining methods reported currently, however, are not satisfactory. The brittle nature of the NiTi/stainless steel (SS) joint was described for the first time by Eijk et al. (2003) in a preliminary study on plasma welding of NiTi and SS. Li et al. (2005) employed capacitor discharge upset welding to join NiTi and SS. The authors reported that while both of the base metals melted, the liquid phase was not removed from the NiTi/SS interface completely, thus resulting in low strength (150 MPa). Fukumoto et al. (2010) used friction welding to prepare NiTi/SS joints, demonstrating that welding without Ni interlayer results in the formation of brittle Fe2Ti phase at the weld interface. This outcome suggested that the solidstate bonding over several seconds failed to inhibit the formation of brittle phases. Gugel et al. (2008) reported the mechanical behaviour of laser-welded NiTi/SS joints, with strengths of up to 623 MPa, but failed to provide any microstructural information.
In order to inhibit the formation of intermetallic phases, Li et al. (2011) introduced Ag–Cu foil as a filler metal in transient liquid phase diffusion bonding (TLPDB). Nevertheless, undesirable phases continued to form in the interfacial regions, rendering the results less than satisfactory. In a comprehensive series of studies on laser welding of NiTi and SS by Li and co-workers, Ni (Li et al. (2012)), Co (Li et al. (2013a)), and Cu (Li et al. (2013b)) were introduced as filler metals. Although the formation of Fe–Ti and Cr–Ti intermetallic phases was inhibited effectively, the excessive addition of filler metals was found to deteriorate the joint properties by inducing proliferation of other brittle phases. The maximum joint strength obtained in the work by Li and co-workers was established as 520 MPa, with 5.1% rupture elongation. In a dissimilar metal weld joint, the existence of brittle intermetallic phases does not necessarily lead to the deterioration of joint properties. For example, in diffusion bonding of Ti to 304 SS, Ghosh et al. (2003) showed that when the growth of the intermetallic compound layer, which consisted mainly of FeTi and Fe2Ti phases, was suppressed to 7 m through temperature optimisation, the tensile strength reached 70% of that of the titanium base metal, with 9% rupture elongation, which indicated good ductility. In dissimilar welding of aluminium alloy to 304 SS, Murakami et al. (2003) showed that when the thickness of the intermetallic layer was decreased to nanometre size, the joint strength exceeded that of the Al base metal. In resistance welding, the formation of brittle phases can be easily suppressed through parameter optimisation, and the detrimental effects of brittle phases can thus be minimised. In order to obtain NiTi/SS joints with good mechanical properties, the following measures were taken in this work:
1) The detrimental effects of intermetallic phases were minimised through adjustment of the welding current; 2) A SS tube was used to improve the strength and bending performance; 3) Post-welding cold drawing was employed to recover softened heat-affected zones (HAZs). In this study, the chemical compositions, microstructures, and microhardness values of the prepared NiTi/SS joints are examined and discussed, and the effects of welding current and cold drawing on microstructures and mechanical properties are investigated.
2. Experimental procedures 2.1 Materials and sample preparation NiTi (50.9 at.% Ni) and AISI304 SS were purchased from Fort Wayne Metals. The wire and tube sizes are marked in Fig. 1. The as-received NiTi wires were subjected to 45 ± 5% cold drawing, followed by superelastic heat treatment. The as-received SS wires were subjected to 93% cold drawing, followed by stressrelieving annealing. All wires were sanded to a bevelled tip. Oxide coatings were removed by electropolishing in a mixed solution of perchloric acid and acetic acid (in a volume ratio of 1:10) at a voltage of 22 V. The wires were ultrasonically cleaned in an acetone bath for 5 min and dried in atmosphere prior to welding.
Fig. 1. A snapshot (left) of the joint assembly and corresponding schematic cross-section (right). The red line in the schematic representation denotes the electrical route. The two bevelled-tip wires are inserted symmetrically into an SS tube, which is surrounded with compacted Al2O3 powder for cooling and constraining purposes. Powder compacting is performed using a custom-built device. Copper electrodes are installed at the end of the SS tube, yielding a sandwich-like assembly. The sandwich-like assembly is mounted onto a custom-made welding mould. The wires in this mould are clamped in order to permit the application of upset pressure.
2.2 Resistance welding As shown in Fig. 1, NiTi and SS wires were inserted into an SS tube at an equal offset, in a manner that enabled their bevelled tips to come into contact with each other. The tube-wire assembly was mounted onto a custom-made device, with which Al2O3 powder was added around the SS tube and compacted under a force of around 1000 N. A copper electrode was installed at both ends of the SS tube after the powder compacting. The sandwich-like assembly was subsequently mounted onto a custom-designed welding mould. The wires in this mould were fixed with clamps and the copper electrodes were connected to a power supply. Upset pressure was applied through the clamps (maximum value of 500 MPa). Resistance welding was achieved by passing a single electrical pulse along the route illustrated in Fig. 1 (red line). The welding experiments were divided into three groups according to the current employed, i.e. 37 A, 40 A and 45 A, and were conducted in a vacuum chamber. Current duration was set to 100 ms and the upset pressure to 270 MPa, i.e. values above which the wire in the unconstrained region (the region between the wire clamp and the copper electrode) would deflect. The 40 A joints (i.e. joints produced at a current of 40 A) were cold drawn from 0.43 mm to 0.36 mm after welding, resulting in a 30% area reduction. For the 45 A joint, however, the drawing process typically resulted in failure because the material was too brittle. For the 37 A joint, only several welding spots were observed at the interface. As a result of a low interfacial strength, NiTi was frequently pulled out from the tube. Microstructural and mechanical information about this type of joint is provided in the supplementary information.
2.3 Metallography and characterisation To analyse the microstructures of the polished joint specimens, NiTi was etched using a mixture of 10% HF, 40% HNO3 and 50% H2O, and SS was etched in a mixture reagent of HCl and HNO3 in a ratio of 3:1. Optical microstructures were visualised using a Zeiss LAB1 microscope. To analyse the weld microstructures, the polished samples were examined by scanning electron microscopy (SEM, JEM2012-FEF) in the backscattered mode. The chemical composition analysis was performed using energy-dispersive spectroscopy (EDS). Microhardness measurements were performed using a Vickers hardness tester with a load of 200 g and a dwell time of 10 s. Fig. 2 shows a schematic representation of the performed hardness distribution measurements. The mean hardness of the base metals was calculated from values measured at five different locations. Static tensile tests were performed using a MTS810 material testing system. In these experiments, the cross-head displacement and the sample gauge length were set to 0.3 mm/min and 50 mm, respectively. The tensile stress in the strain-stress curves was calculated using the external diameter of the SS tube, i.e. 0.43 mm for the as-welded joints and 0.36 mm for the as-drawn joints. The tensile strength was determined as the average of five joints per condition. Finally, SEM was used to examine the fracture morphology.
Fig. 2. Schematic representation of the hardness distribution measurement for each joint. 3. Results and discussion 3.1 Chemical compositions and SEM microstructures of the as-welded joints Fig. 3 shows the SEM micrographs of the 40 A joint obtained in the backscattered mode. It can be observed from Fig. 3b that with the exception of the sharp corners, metallurgical bonding formed at the bevelled NiTi/SS interface. Under upset pressure, the NiTi material fused with the SS tube, as shown in Fig. 3e. As a result of the temperature gradient induced by copper electrodes, the weld shown in Fig. 3c varied in thickness (~1–10 μm) along the bevelled interface. The highest temperature was close to the central point of the bevelled interface, thus making the weld closest to this location the thickest. As shown in Fig. 3c–e, the NiTi/SS weld consisted of dark-grey grains, which exhibited predominantly worm-like morphology, and a light-grey matrix. A subtle gap (~1 μm in thickness) was observed in zone C, as shown in Fig. 3f, indicating insufficient diffusion. The results of EDS analysis of the 40 A joint are shown in Fig. 4. It is apparent from Fig. 4a that in zone A, the content distribution of Fe displays the same trend as the distribution of Cr, and the inverse trend relative to Ni. Ti distribution did not exhibit any apparent correlation with the content distributions of other
components. In spite of the content fluctuation, the peaks in the element composition profiles remained mostly stable, and similar behaviour was observed for the valleys. This alternating pattern suggested that the dark grains and the light-grey matrix arise from two different phases. The element composition profiles in zone C changed smoothly and rapidly across the weld, with no major fluctuations observed in Fig. 4b, which indicates that atomic diffusion has taken place.
Fig. 3. Backscattered SEM micrographs of the 40 A joint. Contrast enhancing was applied to images in (c–d) using Phtotoshop. The image in (a) shows the outline of the joint while (b) focuses on the weld, with magnified views of zones A, B and C provided in (c), (e) and (f), respectively. The magnified view in (d) shows the microstructure of the weld. EDS point analysis was carried
out on the worm-like phase (point 1 in (d)) and inter-grain light-grey matrix (point 2 in (d)).
EDS point analysis was carried out on points marked 1–4 in Fig. 3. Chemical compositions were stamped on the 1000 ºC isothermal section of the Fe–Ni–Ti ternary system, as reported by Cacciamani et al. (2006) and illustrated in Fig. 4c. Quantitative determination of temperature in the current work was difficult because of the very small size of specimens and the closed nature of the welding mould. In solid-state resistance welding, the temperature usually needs to reach ~80–90% of the melting point of the base metal in order to achieve metallurgical bonding. The isothermal section at 1000 ºC was therefore selected in this work. The high cooling rate, however, did not allow sufficient time for atomic diffusion during cooling. As a consequence, the phases formed during the heating stage were likely to be retained at room temperature, regardless of the influence of cooling. Among the main elements detected in the two base metals, the content of Cr was the lowest. For this reason, Cr was expected to have minimal influence on the chemical composition and phase formation of the weld. Therefore, only Ti, Fe and Ni were considered in phase deduction. According to the diffusion path theory suggested by Kirkaldy and Brown (1963), the phases that can be produced in welds are those present in the phase domains crossed by the line connecting the two base metals (Fig. 4c, blue line). As shown in Fig. 4c, therefore, the potential phases produced in the NiTi/SS weld in the present scenario are (Fe, Ni) Ti, Ni3Ti, Fe2Ti and γ-(Fe, Ni). According to the Cr–Ti–Ni ternary phase diagrams reported by Tan et al. (2007) and Cr–Ti–Fe ternary phase diagrams provided by Wang et al. (2017), the solubility of Cr is high in Fe2Ti (>60 at.%) but low in Ni3Ti (11.4 at.%) and NiTi (14.5 at.%). Taking these facts into consideration, the Cr in the NiTi/SS weld was thought to dissolve in Fe2Ti, and the Cr content was linked to the Fe content. This hypothesis was supported by the fact that the Fe and Cr content profiles displayed the same trends in behaviour (Fig. 4a).
Fig. 4. EDS line-scan results obtained for zone A (a) and zone C (b). (c) EDS analysis of the points marked in Fig. 3, stamped on the 1000 ºC isothermal section of a Fe–Ni–Ti phase diagram according to a procedure described by Cacciamani et al. (2006). The blue line in the isothermal section connects the two base metals, and indicates the phases that are potentially present in NiTi/SS joints. According to the diffusion path theory reported by Kirkaldy and Brown (1963), the possible phases are those present in the phase domains crossed by the blue line connecting the two base metals. In addition to the main phases from the base metals, the intermetallic phases that can be present in the joint are thus Fe2Ti and Ni3Ti.
It is clear from Fig. 4c that point 1, which corresponds to the worm-like phase shown in Fig. 3d, was most likely comprised of Fe2Ti, while point 2, corresponding to the light grey matrix in Fig. 3d, was probably a eutectic comprised of (Fe, Ni)Ti/Ni3Ti/Fe2Ti. Points 3 and 4 in zone C corresponded most likely to γ-(Fe, Ni) and (Fe, Ni)Ti, respectively. The worm-like morphology suggested that the weld experienced liquation. The melting point of the weld was probably lowered by the complexity of its chemical composition. According to
the reaction scheme described by Cacciamani et al. (2006) for the Fe–Ni–Ti system, worm-like Fe2Ti was probably produced as the proeutectic phase in the (Fe, Ni)Ti/Ni3Ti/Fe2Ti eutectic reaction. As a result of the growth of worm-like Fe2Ti, the chemical composition of the inter-grain region reached the eutectic point gradually, thereby producing the ternary eutectic. Backscattered SEM micrographs of the 45 A joint are shown in Fig. 5. The bevelled interface appeared steeper than that shown in Fig. 3 for the 40 A joint, most likely as a result of the differences in the cutting angle during sanding. The weld widened to around 40 μm, as shown in Fig. 5c. Under upset pressure, a fraction of mashed NiTi flowed away, thereby leaving behind a hole, as indicated by the yellow ellipse in Fig. 5b. The SS wire-tube reaction layer (RL) in Fig. 5f was found to be rich in Ti, as evidenced by EDS analysis, which indicates that the mashed NiTi was squeezed into the wire-tube gap and reacted with the SS. Fig. 5c shows that the bevelled weld varied slightly in thickness. A magnified view provided in Fig. 5d shows that the weld consisted of two layers, i.e. a RL adjacent to the SS and a NiTi molten zone (MZ). The microstructure of the RL was similar to that observed for the 40 A weld. The dark-grey grains in RL exhibited both equiaxed and needle-like morphologies. Coarse needle-like grains dominated the RL centre, whereas fine equiaxed grains dominated the RL periphery. The NiTi MZ displayed typical melted and re-solidified microstructures, consisting of equiaxed grains and a light-grey matrix. As shown in Fig. 5e, the weld between NiTi and the SS tube also displayed a two-layer structure. The grains in MZ, however, were columnar, an observation that can be attributed to the steep temperature gradient imposed by the heat sink. Grain boundary (GB) liquation was observed in NiTi HAZ (Fig. 5e). An arrested micro-crack, caused most likely by solidification cracking along the columnar grain boundary, was observed near the liquated GBs.
Fig. 5. Backscattered SEM micrographs of the 45 A joint. Contrast enhancing was applied to images in (c–f) using Photoshop. Image in (b) shows a full view of the weld, with zones A, B and C shown at higher magnification in (c), (e) and (f), respectively. The hole highlighted in (b) was produced by mashed NiTi flowing away to the side of the SS, as indicated by the yellow arrow. The magnified view in (d) shows the two-layered structure of the weld, i.e. a reaction layer (RL) adjacent to SS and a NiTi molten zone (MZ). EDS analysis was carried out on the RL (point 1-7), the RL-MZ interfacial region (point 9-8) and the MZ (point 10-15). Grain boundary (GB) liquation was captured in (e).
Fig. 6. Element composition profiles (a) and chemical compositions stamped on phase diagram (b) across the weld. The phase diagram represents the 1000 °C isothermal section of the Fe–Ni– Ti ternary system, as described by Cacciamani et al. (2006). Point 1-7 represents the RL; point 8-9 represents the RL-MZ interfacial region; and point 10-15 represents the MZ
Element composition profiles of the 45 A joint are shown in Fig. 6a. In this system, the Ti distribution displayed major fluctuations over the entire MZ. Further, the distributions of Ni and Fe fluctuated significantly in the interfacial region between RL and MZ. The Cr distribution, by contrast, did not exhibit any significant fluctuations. As shown in Fig. 6b, the chemical compositions of points 1–15 marked in Fig. 5d were stamped on the 1000 °C isothermal section of the Fe–Ni–Ti ternary system. Points 1–7 were in the RL, points 8–9 in the RL–MZ interfacial region, and points 10–15 in the MZ. It is apparent from Fig. 6b that the chemical composition of the NiTi MZ was basically comprised of a Ni3Ti/(Fe, Ni)Ti (a solid solution of NiTi B2 phase, in which Fe atoms substitute a portion of Ni atoms) two-phase domain. Although it is possible that Fe2Ti might exist in the MZ, its presence would represent only a minor phase since the chemical composition of MZ was considerably different from that of the Fe2Ti single-phase domain. Taking into consideration this comparison of chemical compositions, a mechanism for the phase formation in the MZ can be therefore proposed: when molten NiTi starts to solidify, the (Fe, Ni)Ti phase, corresponding to the dark-grey grains in the MZ, grows epitaxially on the grains of the base metal. Accompanying the solidification, the liquid phase was expelled to the inter-grain region, and its chemical composition gradually reached the eutectic point, ultimately giving rise to (Fe, Ni)Ti/Ni3Ti eutectic in the inter-grain region. In the RL, points 1 and 2, corresponding to the light-grey matrix, represent most likely the γ-(Fe, Ni)/Fe2Ti eutectic. Chemical compositions of points 3 and 4 (dark-grey grains in the RL) were comprised of the γ-(Fe, Ni)/Fe2Ti/Ni3Ti three-phase domain, with Ni3Ti and γ-(Fe, Ni) as minor phases. According to the chemical compositions of
points 3–7, the dark-grey grains in the RL were probably composed of Fe2Ti. In the RL–ML interfacial region, the light-grey matrix, corresponding to points 8 and 9, was probably comprised of the Fe2Ti/(Fe, Ni)Ti/Ni3Ti eutectic.
3.2 Metallographic microstructures Fig. 7 shows the optical microstructures of the as-welded 40 A joint. Fig. 7a shows that a gap was produced at the interface as a result of preferential etching. In terms of the SS, although equiaxed grains (~10–20 μm) were prominent near the weld (Fig. 7b), elongated fibre-like microstructures were observed in the base metal (Fig. 7d). The grains shown in Fig. 7c were found to be finer than those in Fig. 7b owing to the lower heating temperature. A heating boundary can be identified in Fig. 7d. In this case, the nearby grains were so fine that it was not possible to identify them using optical microscopy. The as-received SS wires were subjected to 93% cold drawing and subsequent stress-relieving annealing at 350–427 °C. The existence of elongated microstructure in the SS base metal indicates that the stress relieving annealing changed the subgrain structures through dislocation annihilation and polygonisation. According to Naghizadeh and Mirzadeh (2016), the microstructures of the SS base metal consist mainly of stress-induced martensite, and will experience the following processes during heating: 1) reversion of stress-induced martensite to austenite, 2) recrystallisation of retained austenite and 3) growth of the reversed and recrystallised austenite grains. It can be seen from Fig. 7d that the elongated microstructures on the right-hand-side of the heating boundary disappeared, indicating that the reversion process has started. Closer to the weld (Fig. 7c), fine equiaxed grains (~4 m) were observed, which suggests that the reversion and recrystallisation processes have finished. Near the weld, the reversed and the recrystallised austenite grains grew to a length of ~10–20 m, as shown in Fig. 7b.
Fig. 7. Optical microstructures of the as-welded 40 A joint: full view (a). The zones marked in (a) are shown in more detail in the correspondingly labelled images. Images shown in (d) and (i) were captured near the end of the SS tube (these locations could not be included in the image shown in (a)).
In terms of NiTi, the grains were found to be ~10 m in size near the weld, as shown in Fig. 7e. Finer grains were also observed and these are portrayed in Fig. 7g–h. The NiTi base metal did not display any grains when examined by optical microscopy (Fig. 7i). According to the microstructural report provided for NiTi by the manufacturer (Schaffer (2009)), the as-received NiTi wires, which were subjected to 45 5% cold drawing and super-elastic annealing, consisted of nano-sized recrystallised grains (~5–60 nm). The report also stated that when the nano-grains increase in size to 10 μm, the NiTi is expected to lose its superelasticity. During welding, the grains in the NiTi HAZ increased in size, resulting in functional degradation and decline in strength. The presence of lath microstructure near the weld (Fig. 7e–f) indicates the existence of martensite, which was probably caused by the precipitation of Ni-rich phases. In this system, martensite coexisted with the B2 phase near the weld.
Fig. 8. Optical microstructures of the 45 A joint: full view (a), the bevelled weld (b–c), reaction layer (RL) between SS wire and SS tube (d), the weld between NiTi and SS tube (e), NiTi HAZ (f) and SS HAZ (g). The RL and the molten zone (MZ) in images (b) and (e) correspond to those shown in Fig. 5d. The image in (c) shows the solidification mode transition in NiTi MZ. Image in (i) reveals grain boundary (GB) liquation in the NiTi HAZ.
The optical microstructures of the 45 A joint are shown in Fig. 8. As shown in Fig. 8b, it was possible to identify both the RL and MZ in the etched weld. Several holes were observed in the RL as a result of preferential etching, as shown in Fig. 8a and Fig. 8c. The dark-etched RL eutectic shown in Fig. 8b was in good agreement with the results of EDS analysis discussed in the previous section. Fig. 8c shows that the coarseness of the grains increased to ~20 m near the NiTi fusion line. The solidified grains in the MZ grew epitaxially on the base metal, indicating that they had the same crystal structure as the NiTi base metal itself. This observation further corroborates the phase formation mechanism proposed in the previous section. Analyses of the chemical compositions revealed that the solidified equiaxed grains were comprised of (Fe, Ni)Ti (i.e. a solid solution of B2 phase). The (Fe, Ni)Ti grains grew first epitaxially from the fusion line, and in the planar mode along the easy-growth direction of the base metal grains. A short distance away from the fusion line, however, solidification changed to the cellular mode. Deeper into the weld centre, equiaxed dendrites nucleated and grew. This change in solidification mode was caused by increasing constitutional
supercooling in the direction from the fusion line to the weld centre. Fig. 8e shows that the MZ was almost twice as thick as the RL, and the solidification mode changed from planar to cellular, columnar dendritic and equiaxed dendritic. Fig. 8f displays GB liquation that was observed within ~100 μm of the weld, which was caused by the direct melting of the matrix since there was no pre-existing eutectic with a low melting point in the NiTi. According to Kou (2003), the large cooling rate during the welding process can delay the solidification of melted GBs until the eutectic temperature is reached, thereby leading to the formation of an inter-grain eutectic. In the present case, (Fe, Ni)Ti/Ni3Ti eutectic was probably produced after GB solidification. In the SS HAZ (Fig. 8g–h), the grains did not exhibit a further increase in coarseness when compared to those observed in the as-welded 40 A joint. This observation indicates that even though the temperature increased, the welding time was too short for further coarsening. The full RL shown in Fig. 8d reveals a dark-etched eutectic.
Fig. 9. Deformed SS (a) and NiTi (c) at either end of the SS tube, and microstructures (b, d–f) of the as-drawn 40 A joint.
Fig. 9 shows the optical microstructures and deformed wires of the as-drawn 40 A joint. The 40 A joints were cold drawn from 0.43 mm to 0.36 mm, which corresponds to a 30% area reduction. The slope of the bevelled NiTi/SS interface decreased after cold drawing, as shown in Fig. 9b. Fig. 9a shows that no evident microstructural changes were observed at the onset of SS deformation. By contrast, the deformation onset resulted in the formation of a dark boundary on the NiTi side (Fig. 9c). A magnified image shown in Fig. 9f reveals that the dark boundary consisted of a needle-like phase. At this stage, however, it was unclear as to what the needle-like phase was comprised of and why it formed. Fig. 9d shows that the austenite grains in SS HAZ decreased in size to ~5 μm as a consequence of cold drawing and the grain boundaries became more obscure. In terms of NiTi, the grains in the HAZ decreased in size to less than 1μm and became unidentifiable by optical microscopy, as shown in Fig. 9e. In fact, only a scattered needle-like phase was observed
throughout the NiTi HAZ. Microstructure refinement suggested that although the cold drawing resulted in partial recovery of the softened HAZ, further plastic deformation would be required for complete recovery.
3.3 Microhardness and mechanical properties Fig. 10 shows the microhardness distributions and Vickers indentations of NiTi/SS joints. The values of mean hardness for the NiTi and SS base metals were determined as 405 HV0.2 and 578 HV0.2, respectively. For the as-welded 40 A joint, the mean hardness throughout the entire NiTi HAZ was 302 HV0.2. Hardness in the SS HAZ remained stable (189 HV0.2 on average) near the weld, and increased subsequently (in direction away from the weld) to the value of the base metal. Minimum hardness of the joint was established as 172 HV0.2, and existed on the SS side. The hardness profile of the as-welded 40 A joint did not exhibit the pattern characteristic for interlayer-free NiTi/SS joints obtained in previous investigations by Li et al. (2013a), Li et al. (2013b), Li et al. (2012) and Ng et al. (2015), in which a high-hardness plateau was observed in the hardness profile of the welds. As shown in Fig. 10b, the hardness at the NiTi/SS interface was somewhere in between those of the two base metals. After cold drawing, a dramatic increase was observed in the hardness of the as-drawn 40 A joint (Fig. 10a). In fact, the NiTi HAZ was found to be almost as hard as the base metal, and the mean hardness of the SS HAZ increased to 377 HV0.2, which represents a two-fold increase relative to the as-welded joint. The increase in hardness observed for the as-drawn joint was fairly consistent with the results of microstructure refinement.
Fig.10. Microhardness profiles of NiTi/SS joints (a) and microstructures showing Vickers indentations (b–e). Image (b) shows the weld of the as-welded 40 A joint. Images in (c–e) show the bevelled weld, reaction layer between SS wire and SS tube and the weld between NiTi and the SS tube of the 45 A joint. All indentations were created using a 200 g load at a dwell time of 10 s. The hardness values associated with the different indentations are marked in each image.
For the 45 A joint, the hardness profile of the SS HAZ was similar to that of the as-welded 40 A joint. The weld hardness of the 45 A joint, however, increased relative to the 40 A joint as a result of the increased proportions of Fe2Ti and the eutectic. Within ~100 m from the weld, the hardness of the NiTi HAZ exceeded
that of the base metal, which was attributed to GB liquation. As can be seen in Fig. 10c, the highest hardness of the joint was on the NiTi fusion line and the value exceeded 800 HV0.2, thus indicating local embrittlement. The microstructure near the NiTi fusion line was exposed to the coarsest grains throughout the entire HAZ, re-solidified GBs and the eutectic in the MZ. Nevertheless, the deformation coordination between grains was impeded by re-solidified GBs and the eutectic in the MZ, thereby resulting in the embrittlement. The hardness of the RL located between the SS wire and the SS tube (Fig. 10d) indicates that the complex eutectic structures did not embrittle the material severely. The hardness on the NiTi fusion line (Fig. 10e) was determined to be only 455 HV0.2, i.e. much lower than that shown in Fig. 10c owing to the presence of finer grains (~10 μm). Owing to the columnar structure, the hardness of the NiTi-tube weld was higher than that of the bevelled weld shown in Fig. 10c.
Fig. 11. Tensile deformation behaviour of dissimilar NiTi/SS joints. Three curves are presented for each condition.
Fig. 11 demonstrates the tensile deformation behaviour of NiTi/SS joints. Average tensile strength and rupture elongation of the as-welded 40 A joint were determined to be 440 MPa and 7.9%, respectively. It is clear from the strain-stress curves that the joints experienced stress plateaus prior to their ultimate fracture. For the 45 A joint, the embrittlement near the NiTi fusion line resulted in the deterioration of mechanical properties, with the mean tensile strength and elongation declining to 340 MPa and 5.8%, respectively. For this joint, the fracture occurred before reaching the maximum strain of the stress plateau. The mean tensile strength of the as-drawn 40 A joints reached 830 MPa, with a slight decrease in the mean fracture elongation to 6.2%. The strain-stress curves of these joints show that major stress fallings occurred just before the stress plateaus, with several minor ones appearing also after the plateaus. This odd deformation behaviour is quite common for resistance-welded joints, and is most likely related to the interactions formed between the peripheral SS tube and the core wires under tensile load. The fracture morphologies of NiTi/SS joints are shown in Fig. 12. The as-welded 40 A joint was found to fracture at the root of the NiTi slope and additionally, the NiTi wire was pulled out from the SS tube (Fig.12a). Fine dimples were observed on the fracture surface shown in Fig. 12b, indicating that the joint was
broken as a result of micro-void coalescence. The side surface of the NiTi wire shown in Fig. 12c exhibited a rough morphology, with many longitudinal ridges. Under tensile load, the ridges acted as microtube-wire bonding devices, thereby enhancing the joint strength. As shown in Fig. 13, the as-welded 40 A joint was bent around a rod with a diameter of 2 mm. Instead of fracturing, however, a plastic deformation was observed after bending, thereby indicating good plasticity and bending performance.
Fig. 12. Fracture morphologies of NiTi/SS joints: the as-welded 40 A joint (a–c), the as-drawn 40 A joint (d–f) and the as-welded 45 A joint (g–i). The focus zones marked in (a), (d) and (g) are shown at higher magnification in the images labelled with corresponding letters.
Fracture morphology of the as-drawn 40 A joint is shown in Fig. 12d–f. The joint fractured at the NiTi slope root as well. Dimples dominated the fracture surface, thus indicating the occurrence of a micro-void coalescence failure. According to the microstructures and microhardness of the as-drawn joint, the cold drawing reinforced the softened HAZs. As a result of the shrinking after cold drawing, the external SS tube was able to secure the core wire more tightly. Overall, these two points improved the tensile strength of the as-drawn joint. As shown in 12g, the 45 A joint cracked on the NiTi side, a process that was accompanied by tube fracture, which could be attributed to the brittle character of the weld between the tube and NiTi. As shown in Fig. 10e, columnar grains grew perpendicularly to the tube wall, resulting in strong anisotropic mechanical
properties. Under tensile load, cracks were typically initiated in the inter-grain region where the eutectic existed, and propagated along the columnar grain boundary, extending to the tube wall and eventually resulting in its breakage. Both cleavage terraces and river patterns were observed on the fracture surfaces, which are indicative of cleavage fracture mechanism. Local embrittlement near the NiTi fusion line in particular, was the primary reason behind the cleavage fracture. The direction of the river patterns suggested that the failure initiated from the crack shown in Fig. 12h.
1 mm
Fig. 13 A bent as-welded 40 A joint. Although the bending test caused a considerable plastic deformation, the joint did not break.
4. Conclusions Herein, resistance welding was employed in order to fabricate NiTi/SS joints for biomedical applications. Effects of welding current and post-welding cold drawing on microstructures and mechanical properties of the joints were investigated. The following conclusions can be drawn from the current study:
1.
The increase in welding current widened the weld and altered its microstructure. At a welding current of 40 A, the weld was comprised of a RL with 10 m in size. At 45 A welding current, the weld widened to 40 m, with a NiTi MZ produced in addition to the RL.
2.
The weld formed between the NiTi wire and the SS tube increased the bonding area. At a welding current of 45 A, the NiTi MZ of the NiTi/tube weld displayed a columnar microstructure with some micro-cracks.
3.
For the 40 A joint, the 30% area reduction induced by the cold drawing caused a decrease in the size of the grains in the SS HAZ to 5 m, and a two-fold increase in microhardness. By contrast, the grains in the NiTi HAZ decreased in size to less than 1 m and the microhardness remained similar to that of the base metal.
4.
The tensile strength and rupture elongation of the 40 A joint were established as 440 MPa and 7.9%, respectively. After cold drawing, the tensile strength increased to 830 MPa, while the
rupture elongation decreased slightly to 6.2%. The tensile strength and rupture elongation of the 45 A joint were determined as 340 MPa and 5.8%, respectively. Acknowledgement Funding: This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.
References Cacciamani, G., De Keyzer, J., Ferro, R., Klotz, U.E., Lacaze, J., Wollants, P., 2006. Critical evaluation of the Fe–Ni, Fe– Ti and Fe–Ni–Ti alloy systems. Intermetallics 14, 1312-1325. Eijk, C.V.D., Fostervoll, H., Sallom, Z.K., Akselsen, O.M., 2003. Plasma welding of NiTi to NiTi, stainless steel and hastelloy C276. Joining of Advanced and Specialty Materials VI, 125-129. Fukumoto, S., Inoue, T., Mizuno, S., Okita, K., Tomita, T., Yamamoto, A., 2010. Friction welding of TiNi alloy to stainless steel using Ni interlayer. Sci Technol Weld Joi 15, 124-130. Ghosh, M., Bhanumurthy, K., Kale, G.B., Krishnan, J., Chatterjee, S., 2003. Diffusion bonding of titanium to 304 stainless steel. Journal of Nuclear Materials 322, 235-241. Gugel, H., Schuermann, a., Theisen, W., 2008. Laser welding of NiTi wires. Materials Science and Engineering: A 481482, 668-671. Kirkaldy, J.S., Brown, L.C., 1963. Diffusion Behaviour in Ternary, Multiphase Systems. Canadian Metallurgical Quarterly 2, 2:89. Kou, S., 2003. Welding Metallurgy, 2nd Edition. 310-346. Li, H., Li, Z.X., Wang, Y.L., Feng, J.C., 2011. Transient Liquid Phase Diffusion Bonding of TiNi Shape Memory Alloy and Stainless Steel. Rare Metal Mat Eng 40, 1382-1386. Li, H., Sun, D., Cai, X., Dong, P., Gu, X., 2013a. Laser welding of TiNi shape memory alloy and stainless steel using Co filler metal. Optics & Laser Technology 45, 453-460. Li, H., Sun, D., Gu, X., Dong, P., Lv, Z., 2013b. Effects of the thickness of Cu filler metal on the microstructure and properties of laser-welded TiNi alloy and stainless steel joint. Materials & Design 50, 342-350. Li, H.M., Sun, D.Q., Cai, X.L., Dong, P., Wang, W.Q., 2012. Laser welding of TiNi shape memory alloy and stainless steel using Ni interlayer. Materials & Design 39, 285-293.
Li, M., Sun, D., Qiu, X., Sun, D., Yin, S., 2005. Microstructures and properties of capacitor discharge welded joint of TiNi shape memory alloy and stainless steel. China Welding, 95-100. Murakami, T., Nakata, K., Tong, H., Ushio, M., 2003. Dissimilar Metal Joining of Aluminum to Steel by MIG Arc Brazing Using Flux Cored Wire. Transactions of the Iron and Steel Institute of Japan 43, 1596-1602. Naghizadeh, M., Mirzadeh, H., 2016. Microstructural Evolutions During Annealing of Plastically Deformed AISI 304 Austenitic Stainless Steel: Martensite Reversion, Grain Refinement, Recrystallization, and Grain Growth. Metall Mater Trans A 47A, 4210-4216. Ng, C.H., Mok, E.S.H., Man, H.C., 2015. Effect of Ta interlayer on laser welding of NiTi to AISI 316L stainless steel. Journal of Materials Processing Technology 226, 69-77. Schaffer, J.E., 2009. Structure-Property Relationships in Conventional and Nanocrystalline NiTi Intermetallic Alloy Wire. J Mater Eng Perform 18, 582-587. Tan, Y.-h., Xu, H.-h., Du, Y., 2007. Isothermal section at 927 °C of Cr-Ni-Ti system. Transactions of Nonferrous Metals Society of China 17, 711-714. Wang, S., Wang, K., Chen, G., Li, Z., Qin, Z., Lu, X., Li, C., 2017. Thermodynamic modeling of Ti-Fe-Cr ternary system. Calphad 56, 160-168.
Figure and Table Captions. 1.
A snapshot (left) of the joint assembly and corresponding schematic cross-section (right).
2.
Schematic of hardness distribution measurement for the joint.
3.
Back scattered micrographs of the 40 A joint.
4.
EDS analysis results of the 40 A joint.
5.
Back scattered micrographs of the 45 A joint.
6.
EDS analysis results of the 45 A joint.
7.
Optical microstructures of the 40 A as-welded joint.
8.
Optical microstructures of the 45 A joint.
9.
Optical microstructures of the 40 A as-drawn joint.
10. Hardness profiles and Vickers indentations of NiTi/SS joints. 11. Tensile performance of NiTi/SS joints.
12. Fracture morphology of NiTi/SS joints. 13. A bent 40 A as-welded joint.
Figr-1
Figr-2
Figr-3
Figr-4
Figr-5
Figr-6
Figr-7
Figr-8
Figr-9
Figr-10
Figr-11
Figr-12
Figr-13