influence of welding pressure on diffusion welded joints ...g. thirunavukarasu, b. murugan, s....

WELDING RESEARCH

FEBRUARY 2017 / WELDING JOURNAL 53-s

Introduction Diffusion welding (DW) is a high-temperature solid-state joiningprocess that permanently joins matingsurfaces by simultaneous applicationof pressure and heat. It does not in-volve macroscopic deformation, melt-ing, or relative motion of the partswelded. A solid filler metal (diffusionaid) may be inserted between the fay-ing surfaces (Ref. 1). DW has demon-strated its uniqueness by finding ap-plication in the processes and productsof front line areas of science and tech-nology. DW is actually gaining mo-

mentum of applicability from the daysof conception due to the inappropri-ateness of other joining techniques. In the past quarter-century, Ti al-loys have been successfully used in air-crafts, rockets, and biomedical implants, among other products, as keystructural materials. Ti alloys derivetheir excellent properties from a combi-nation of two ductile phases (hcp andbcc allotropic modifications) of Ti. Sev-eral key features of these phases inde-pendently and dependently determinethe plasticity and fracture behavior ofthe family of Ti-alloys (Ref. 2). Both thehigh corrosion resistance and the good

strength/weight ratios of titanium andits family of alloys make them an attrac-tive prospect for use in offshore oil andgas platforms. Ti-alloy armor is beingincorporated into the design of somecombat vehicles in order to saveweight (Ref. 3). Although Ti alloys are nonmagnet-ic, their high electrical resistivity isused in the manufacture of electro-magnetic cookware. In addition to thecapability of electromagnetic cooking,titanium’s light weight, low heating ca-pacity, and good corrosion resistanceare the favorable properties of thesecookwares (Ref. 4). Austenitic stainless steels (ASSs)are the most common and familiartype of stainless steels. They are recog-nized as nonmagnetic. They are ex-tremely formable and weldable, andsuccessfully used from cryogenic tored-hot temperatures. ASSs have ex-ceptional toughness and a combina-tion of strength and corrosion resist-ance at elevated temperatures. ASSsare the most creep resistant of thestainless steels. Lean austenitic alloysconstitute the largest portion of allstainkess steels (SSs) produced. Theseare principally 201, 301, and 304grades. Alloys with less than 20% Crand 14% Ni fall into this unofficial cat-egory (Ref. 5). Highly efficient and re-liable joints of corrosion-resisting-steels with titanium alloys find thegreatest inevitable need in nuclearpower engineering (Ref. 6). In a satellite cooling system, Ti-6Al-4V is joined to stainless steel

Influence of Welding Pressure on Diffusion WeldedJoints Using Interlayer

Considering diffusion welding for superior performances

BY G. THIRUNAVUKARASU, B. MURUGAN, S. CHATTERJEE, AND S. KUNDU

ABSTRACT The present study focused on elucidating the extent of influence of welding pressureon the evolution of interfacial microstructure, mechanical properties, and fracture morphologies of solidstate diffusionwelded joints (DWJs) of Ti6Al4V (TiA) and 304stainless steel (SS) using 200μmthick nickel (Ni) as an interlayer processed in vacuum at750°C for 60 min using welding pressure in the range of 2 to 10 MPa in step of 2 MPa.Layer wise Ni–Tibased intermetallic phase(s) were revealed at the TiA|Ni interface,whereas the Ni|SS interface showed solid solution behavior. From the surface propertyof the DWJs, it was sensed that the effect of work hardening occurred when the weldingwas processed at 4 MPa and above welding pressures. Exponential growth in the bulgingproperty of the DWJs with the increase in the welding pressure was spotted. The studyindicated the significant influence of welding pressure over the welded assemblies’ tensile properties; however, tensile properties behaved asymptotic when the DWJs wereprocessed beyond 6MPa welding pressure. To understand the fracture characteristic ofthe DWJs, fracture path and fractography were evaluated. Thermal stresses induced onthe materials diffusion welded were calculated.

KEYWORDS • Diffusion Welding • Ti Alloys • Stainless Steels • Intermetallics • Mechanical Properties

G. THIRUNAVUKARASU, B. MURUGAN, S. CHATTERJEE, and S. KUNDU ([email protected]) are with the Department of Metallurgyand Materials Engineering, Indian Institute of Engineering Science and Technology, Shibpur, India.

Thirunavukarasa - Influence of Welding Pressure _Layout 1 1/16/17 9:28 AM Page 53

WELDING RESEARCH

WELDING JOURNAL / FEBRUARY 2017, VOL. 9654-s

(1Cr18Ni9Ti) (Ref. 7). Joints of Ti al-loy and Type 304 ASS are used in fueltanks of satellite launch vehicles (Ref.8); hence, joints of Ti-based alloy andSS using different welding techniquesare undertaken as sponsored researchby different agencies and institutes toknow the details of optimal levels ofoperating parameters to achieve farbetter performance. Table 1 shows the physical proper-ties of Ti-6Al-4V and 304 SS (Ref. 9).It is readily seen that large propertydifferences exist between the metals.The large variation of these and otherproperties provide an attractive sourcefor the choice of materials for specialservice conditions and environments;however, the large differences in phys-ical properties also yield difficulties injoining these metals together. A largedifference in the melting temperature

between two metalsmakes the conventionalfusion-welding process-es inapplicable due tothe segregation of low-melting eutectics,which can cause hotcracking. Large differ-ences in thermal expan-sion will lead to the for-mation of large residualstresses and this leadsto the formation of mi-crocracks at the weldedregion; and thus reducethe joint strength andcause fatigue problems.A difference in the ther-mal conductivity of twometals easily causes un-even heat dissipation.Furthermore, chemicalmismatches in dissimi-lar-metal joints can re-sult in the formation ofbrittle phases and thediffusion of particularelements. These will ad-versely affect the serv-

ice properties of the joints, and are themain reasons why dissimilar-metaljoining is normally more complex thansimilar-metal joining in many re-spects. Consequently, a solid-statejoining method, viz., DW of these twomaterials using an interlayer, hasgained more importance than otherconventional welding methodologies.During welding, Ti alloys pick up oxy-gen and nitrogen from the atmospherevery easily (Refs. 10, 11). For this rea-son, DW in a vacuum or inert gas at-mosphere is suggested. One of the necessary conditions toconsidering a material as a potentialinterlayer to join the base metal(s) isthat it reduces the chemical hetero-geneity and improves the thermody-namic stability of the diffusion zone atthe interface of the welded joints. Zeeret al. (Ref. 12) proposed a method to

classify interlayers for DW of dissimi-lar materials and discussed the roleand contribution of the interlayers inthe formation of the diffusion weldedjoints (DWJs) in different stages ofwelding. They advised using interlay-ers made from alloys or pure metals,usually with high ductility, such asnickel, copper, cobalt, zirconium,vanadium, Permalloy, silver and gold.The interlayers were used in the formof foil, film, or coatings and depositedon one or both sides of the to-be-weld-ed surfaces by electroplating or ther-mal spraying, rolling of powders, andin the combined form of spraying andfoil. Nickel was considered as a poten-tial candidate for the interlayer due toits corrosion resistance for applicationat high temperature and substantialsolid solubility with iron. Ni–Ti binaryphase diagram suggested that inter-metallics formation changes fromNiTi2 to Ni3Ti via NiTi with increasingNi (Refs. 13, 14). He et al. (Ref. 15) re-ported Ni–Ti intermetallics have high-er plasticity than Fe–Ti intermetallics. The success or failure of the DWprocess depends on welding pressure,temperature, and time (Ref. 16). In thepast, most literature reported eitherthe effect of temperature or time onthe DWJs’ microstructure and me-chanical properties. The present au-thors along with Professor Mishra car-ried out two separate investigationson DWJs of TiA and SS using 200-μm-thick Ni as an interlayer. To first opti-mize the welding temperature (Ref. 17),DWJs were prepared in the temperaturerange from 700° to 800°C in steps of25°C using 3-MPa pressure for 60 min.A maximum tensile strength of ~206MPa along with a ductility of ~2.9% wasobtained for the DWJs processed at750°C. Secondly, to optimize the weld-ing time (Ref. 18), DWJs wereprocessed using 4-MPa pressure at750°C by varying the time from 30 to120 min in steps of 15 min. Maximumtensile strength of ~382 MPa along withelongation of ~3.7% was observed forthe DWJs processed for 60 min. Al-though it is universally accepted thatwelding pressure is a very importantvariable in the solid-state DW process,up until now there existed no sufficient,well-established empirical study/proofto back up such a hypothesis. In view of limited studies, the pres-ent research work described in this pa-

Fig. 1 — Schematic diagram of the indigenous fixture usedfor processing DWJs.

Table 1 — Physical Properties of Base Metals

Metal Melting Density Thermal Conductivity Coefficient of Thermal Point (°C) (g/cc) (W/m–K) Expansion (μm/m–°C), linear 250°C

304 SS 1400–1455 8 16.2 17.8Ti6Al4V 1604–1660 4.43 6.7 9.2


WELDING RESEARCH


per was undertaken by the researchersin an attempt to know the extent of sig-nificant influence of pressure variationover intermetallics formation at the dif-fusion zone, the lateral plastic deforma-tion, the room-temperature mechanicalproperties (tensile properties and hard-ness), and the fracture geographies ofDWJs and to provide information aboutthe possible causes. As this work is incontinuation of the previous works(Refs. 17, 18), the optimized tempera-ture of 750°C and the optimized time of60 min were chosen from the respectiveliterature sources (Refs. 17, 18). Choos-ing higher welding temperature andlonger welding time deteriorated theDWJs’ strength due to the formation ofa large volume of Ni–Ti-based inter-metallics at the TiA|Ni interface (Refs.17, 18). Even higher temperature andlonger time will lead to the formation ofFe–Ti-based intermetallics that aremore detrimental to the mechanicalproperties of the DWJs. Herein, the ex-perimental investigation of influence ofwelding pressure on the evolution of in-terfacial microstructure and the me-chanical properties of solid-state diffu-sion welded joints of Ti-6Al-4V (TiA)and 304 stainless steel (SS) using 200-μm-thick Ni as an interlayer wasprocessed in vacuum at 750°C (weldingtemperature) for 60 min (welding time)using welding pressure in the range of 2to 10 MPa in steps of 2 MPa.

Experimental Procedures

Base metals: Ti-6Al-4V wt-% [a.k.a.TiA] and Fe-18Cr-8Ni wt-% [a.k.a. SS]and interlayer: 99.5 wt-% Ni (~200 ± 10μm), were used in this investigation.Mating surfaces of base metals and in-terlayer were polished using diamondpaste (1 μm) after grinding by hand us-ing eleven series of SiC emery papers(grit sizes 40# to 1600#) to reduce sur-face roughness. Before DW, the matingsurfaces were cleaned with acetone anddried using a portable hair dryer. In thisstudy, DWJs were processed using aspecial fixture (Fig. 1) designed andmanufactured indigenously. The fixturedesign is capable of processing only oneDW experiment at a given point intime; hence, the DW experiments aretime consuming by nature and by de-sign. Mica of ~1-mm thick was kept onthe supporting die (Inconel) of the fix-ture — Fig. 1. The TiA|Ni|SS assemblywas placed over the mica. Mica avoidsthe interfacial contact between the as-sembly and the supporting die. Overthe assembly, mica was again placed toavoid the interfacial welding betweenthe plunger (Inconel) and the assembly.A ceramic-type insulator was used in be-tween load cell and plunger to preventmetal-metal contact. Over the load cell,screw threaded wheel (of stainless steel)for load application was mounted. After

positioning all the above items in-line, aload cell (capacity: 500 kg) was connect-ed to the load indicator, and the re-quired welding pressure (2–10 MPa)was applied along the centerline of theassembly using the threaded wheel thatacts over the load cell and the to-be-dif-fusion-welded assembly. This threadedwheel is used to control the weldingpressure applied on the assembly. Thewelding pressure applied was measuredat room temperature. The fixture withthe assembly was inserted inside a vacu-um chamber, with one end of the vacu-um chamber sealed using nuts and boltsand the other end connected to a vacu-um unit. After reaching enough vacuum(~6–8 10-3 Pa), temperature was in-creased from 32C (room temperature)to 750C (welding temperature) at therate of 0.24Cs-1. The TiA|Ni|SS assem-blies were held for 60 min (welding in-terval) at 750C. Power supply to thewelding furnace was switched off after60 min. At each level of the weldingpressure factor, three separate experi-ments were conducted to process threesamples of DWJs. In this study, experiments wereperformed at five different levels (2, 4,6, 8, and 10 MPa) of welding pressurefactor; hence, there are a total of 15 in-dividual, independent, and random ex-periments (three experiments repeat-ed at each level of the welding-pres-sure factor). After the DW process, DWJs weresectioned perpendicular to the matingsurface and the surfaces were preparedby metallurgical polishing techniques.Polished surfaces of DWJs were exam-ined in an SEM (JSM-5510, JEOL) inbackscattered electron mode (SEM-BSE) to obtain finer microstructuralfeatures of diffusion zone. Chemicalcomposition of the reaction layerswere determined by energy dispersivespectroscopy (SEM-EDS, Thermo Elec-tron Corporation-Noran System SixC10018) using a SiLi detector. Tensileproperties of DWJs were evaluatedwith an Instron 4204 at a cross-headspeed of 8.33 10-4 mms-1. Cylindricaltensile specimens (gauge diameter: 4mm, gauge length: 20 mm) were machined as per ASTM specificationE8M-11 (Ref. 19). The average of the tensile properties of three DWJsprocessed at each pressure level alongwith the associated errors were report-ed here.

Fig. 2 — SEMBSE images of TiA|Ni interface of DWJs processed at 750C for 60 minusing A — 2 MPa; B — 4 MPa; C — 6 MPa; D — 8 MP.

A

C D

B

Thirunavukarasa - Influence of Welding Pressure _Layout 1 1/12/17 5:27 PM Page 55

WELDING RESEARCH


Microhardness measurements werecarried out on the polished surface ofDWJs using Vickers diamond indenter(Leica VMHT) with a load of 25 gramforce for 20 s dwelling time. Five meas-urements were carried out at each sec-tion and the respective average wasmentioned in this study. Fracture sur-faces of DWJs were observed in second-ary electron mode using the SEM-EDSto reveal the nature and most possiblelocation through which failure was initi-ated during tensile loading. SEM-BSEobservations were also made to studythe location of the fracture path, whilefractography examination was carriedout using secondary electrons of SEM.

Results Absence of discontinuities and voidsat both the interfaces (both TiA|Ni and

Ni|SS in Figs. 2, 3, respectively) was re-vealed, which established the superiorcontact of the mating surfaces withprocess parameters; hence, DWJs wereefficacious. Layer wise intermetallicswere observed at the TiA|Ni interface(Fig. 2), whereas the Ni|SS interface(Fig. 3) revealed solid-solution behavior.It was observed that the TiA|Ni inter-faces (Fig. 2) of the DWJs processed us-ing 6 and 8 MPa had only a light-shadedreaction layer. The light-shaded reactionlayer on an average consisted of Ti (~27at.-%) and Ni (~70 at.-%); hence, theNi–Ti binary phase diagram indicatedthe realization of Ni3Ti intermetallic(Refs. 13, 14). The TiA|Ni interface of the DWJsprocessed using 2, 4, and 10 MPa hadtwo reaction layers. The deep-shaded re-action layer observed adjacent to theTiA matrix consisted of Ti (~50 at.-%)

and Ni (~45 at.-%); hence, the Ni–Tiphase diagram indicated the possibilityof realization of NiTi intermetallic(Refs. 13, 14). The reaction layer adja-cent to the Ni matrix was light shadedNi3Ti intermetallic. In general, for allDWJs, the width of Ni3Ti intermetallicwas wider compared to the NiTi inter-metallic at the reaction zone (Ref. 18).The formation of the similar Ni–Ti-based reaction products were also ob-served in the literature (Refs. 20, 21). It was observed, the hardness ofthe TiA side (Fig. 4) of the DWJs was~400 HV irrespective of the weldingpressure adopted during the experi-ments. The hardness of the other basemetal side (SS side) of the DWJs waswitnessed to be ~300 HV for theDWJs processed at 2 MPa, whereasthe surface hardness of the SS side was~400 HV for the DWJs processed from

Fig. 3 — SEMBSE images of Ni|SS of DWJs processed using welding pressure of A — 2 MPa; B — 10 MPa.

Fig. 4 — Microhardness profile along the interfaces of DWJsprocessed using various welding pressures.

Fig. 5 — Variation of lateral deformation (%) in TiA and SS ofDWJs with respect to the welding pressure applied.

A B

Thirunavukarasa - Influence of Welding Pressure _Layout 1 1/12/17 5:27 PM Page 56

WELDING RESEARCH


4 to 10 MPa. In the same way, thehardness of the interlayer zone (Ni) ofthe DWJs was found to be ~175 HVfor the DWJs processed at 2 MPa,whereas it was ~200 HV for the DWJsprocessed from 4 to 10 MPa. It was ob-served that the increase in the weldingpressure had a significant influence inincreasing the bulk property (hard-ness) of the SS side and Ni zone butsuch phenomenon was not observedin the case of the TiA-side. The materials that were diffusionwelded underwent serious changes inshape, i.e., geometry. In reality, when-ever a force is applied to a solid object,a stress is produced within the object.The physical effect/manifestation ofthis stress is that the object experi-ences some form and amount of defor-mation, a strain. As the DWJs wereprepared at five different levels of thewelding-pressure factor, it was decidedto observe the lateral deformation ofthe DWJs. The lateral deformation(Equation 1) of the DWJs is defined asfollows:

d1: (diameter of base metal section) be-fore diffusion-bonding

d2: (diameter of base metal section) af-ter diffusion-bonding

The details of the lateral deformationof TiA and SS are given in Fig. 5. Thelateral deformation of the TiA side ofthe DWJs was much higher than that

of the lateral defor-mation of the SSside of the respec-tive DWJs. The rea-son for the signifi-cant differences inthe lateral deforma-tion of TiA side andSS side needs to befigured out. Both tensilestrength and elonga-tion (Fig. 6) in-creased with the in-crease in the weldingpressure from 2 to 6MPa. Increase in thestrength and elongation of the DWJswith the increase in pressure was pre-sumably due to the improvement in thecoalescence of the mating surfaces ofthe DWJs. There was no significant dif-ference in the tensile strength of DWJsprocessed above 6 MPa. The tensilecurve was more or less flat for theDWJs processed above 6 MPa. Elonga-tion dropped insignificantly for theDWJs processed above 6 MPa. A slightdrop in the elongation of DWJs withthe increase in the welding pressurefrom 6 to 10 MPa is due to the workhardening effect (Ref. 22) experiencedby DWJs. DWJs processed at 6 MPaachieved the optimal strength of ~560MPa with substantial elongation of ~12%. In this work, dissimilar metal jointsof Ti-6Al-4V and 304 SS achieved thehighest strength of ~560 MPa, which isalso the best performance of the jointsprocessed to date in other routes re-ported in literature (Refs. 23–29).Shanmugarajan et al. (Ref. 23) achieved

~40 MPa for the laser welded joints ofTi and SS using Ta as an interlayer.Wang et al. (Ref. 24) were able to reach~124 and ~234 MPa for the electronbeam welding of titanium alloy to SSusing Ni and Cu filler metal, respective-ly. Balasubramanian (Refs. 25, 26)achieved ~158 MPa for DWJs of Ti-6Al-4V and 304 SS using an Ag interlayer.Tomashchuk et al. (Ref. 27) reached~350 MPa for electron beam weldedjoints of Ti alloy and SS via copper asinterlayer. Norouzi et al. (Ref. 28)achieved ~374 MPa for transient liquidphase (TLP) welded joints of Ti-6Al-4Vand AISI 304 austenitic SS using Cu in-terlayer. Kumar et al. (Ref. 29) accom-plished the highest tensile strength of~523 MPa for the dissimilar metaljoints of Ti-6Al-4V and SS304L with Cuas interlayer through friction weldingtechnique. This work proved DW is apromising route to process dissimilarmetal joints with far better mechanicalproperties.

Lateral deformation (%)

= (d2 �d1 )d1

� 100 (1)

Fig. 6 — Tensile strength (MPa) and breaking strain (%) of DWJsprocessed at different levels of welding pressure.

Fig. 7 — Fracture path of DWJs processed at A — 2 MPa—SSside; B — 6 MPa—SS side.

A

B

Thirunavukarasa - Influence of Welding Pressure _Layout 1 1/12/17 5:27 PM 7

WELDING RESEARCH


Metallic mounts were preparedfrom the SS side of the tensile-fractured DWJs. The SS side (Fig. 7A)of the fractured DWJ processed at 2MPa revealed ample Ni interlayer;hence, the SEM micrographs of themetallic mounts suggested that thelower welding pressure functionallyfavored the fracture along the TiA|Niinterface with the intermetallics(Ni3Ti and NiTi) completely cleavedoff the TiA matrix. The fracture path(Fig. 7B) of the DWJ processed using6 MPa exposed the full length of theNi layer. It was observed that theDWJ processed using 6 MPa hadproofs of Ni3Ti intermetallic attachedto the edge of Ni foil; therefore, theelectron micrographs of the metallicmounts recommended that DWJsprocessed at higher welding pressurealso preferred to fracture along theTiA|Ni interface with the intermetal-lic (Ni3Ti) completely hewed off theTiA matrix. In brief, the fracture path(Fig. 7) observation yielded valuableinformation that the TiA|Ni|SSDWJs fractured along the TiA|Ni in-

terface regardless of the welding pres-sure chosen. Selective tensile fractured samplesalong the TiA side and SS side of DWJsprocessed at 2 and 8 MPa were shownin Fig. 8. Both the fracture features ofthe TiA side (Fig. 8A) and SS side (Fig.8B) of DWJ processed at 2 MPa didnot show any sign of dimpled-rupture(ductile) mode of failure; but showedthe noted “river marks.” Both the frac-tographs (Figs. 8A and B) had typicalslightly faceted appearances principal-ly because of the changes in the orien-tation of the cleavage planes operatingfrom grain to grain; hence, it was es-tablished that the DWJs processed at2 MPa had undergone a majorly brittlefracture of intergranular mode appar-ently through the TiA|Ni interfaceduring tensile loading. Both the fracture structures of theTiA side (Fig. 8C) and SS side (Fig. 8D)of the DWJ processed at 8 MPa detect-ed two distinct regions. Firstly, the

zones identified as white circles werefaceted river-pattern rupture (brittle)mode of failure; secondly, the zonesidentified using black rectangles weredimpled-rupture (ductile) mode of fail-ure. Although the faceted river pat-terns were more in volume as com-pared to the other feature in the frac-tographs (Figs. 8C and D), it could becomfortably claimed that the DWJsprocessed using 8 MPa had undergonea mixed mode of failure. For better vi-sualization and understanding, the de-tails of the zone identified as whitecircles in Figs. 8C and D were shown inFigs. 8E and G, respectively. Similarly,the details of the zone identified asblack rectangles in Figs. 8C and D wereshown in Figs. 8F and H, respectively.In general, it was learned that the low-est welding pressure (2 MPa) yieldedbrittle fracture morphologies; and fi-nally, the higher welding pressure (8MPa) yielded mixed mode fracturalfeatures.

Fig 8 — Fracture surface of DWJs processedat 750°C for 60 min using A — 2 MPa, TiAside; B — 2 MPa, SS side; C — 8 MPa, TiAside; D — 8 MPa, SS side; E — 8 MPa, TiAside, white circled in Fig. 8C; F — 8 MPa,TiA side, black rectangled in Fig. 8C; G — 8MPa, SS side, white circled in Fig. 8D; H — 8MPa, SS side, black rectangled in Fig. 8D.

F

A

D E

HG

B C


WELDING RESEARCH


Discussion Sensible reasons for the presence ofthe layer-wise Ni–Ti-based inter-metallics at the TiA|Ni interface andthe solid-solution behavior at theNi|SS interface could be explained us-ing Hume-Rothery rules and the prin-ciple of electronegativity (Ref. 18). Atthe interfaces, the atoms from bothsides diffuse in the opposite direction,but there is an effective diffusion ofatoms at each interface that is morepredominant than the other direction.At the TiA|Ni interface, Ni atoms dif-fused deeper into the TiA matrix ascompared to the diffusing capability ofTi atoms into the Ni matrix. Similarly,net flux of Ni atoms past the interfacein one direction were faster than theflux of atoms in SS in the opposite di-rection; hence, the effect proposed byKirkendall was observed at Ni|SS in-terface for all DWJs (Ref. 18). As DWJs fractured at the TiA|Ni in-terface due to the presence of inter-metallics at the interfacial reaction (dif-fusion) zone, the diffusion zone width(DZW) of the TiA|Ni interface wasmeasured. The DZW (Fig. 9) at the in-terface of the DWJs processed at 2 and4 MPa have negligible significant differ-ences. The DZW decreased with the in-crease in the welding pressure from 4 to8 MPa. The DZW increased with the in-crease in the welding pressure from 8 to10 MPa. A decrease in the DZW with anincrease in the welding pressure, from 4to 8 MPa, indicated there was a decreasein the net activity (diffusivity) of atomsat the TiA|Ni interfaces. A decrease in

the net activityof atoms withan increase inwelding pres-sure showsan inverse re-lationship. Most ofour engineering materials are poly-crystalline so that grain boundary dif-fusion is far more important than sur-face diffusion. Deformation forces thedislocations that exist in a polycrys-talline metal to move up and pile up atthe grain boundaries, and new disloca-tions to be generated. Consequently,the increase in pressure hinders thegrain boundary diffusion, which af-fects the surface diffusion phenome-non. Hence, interfacial activity was af-fected (Refs. 30–32). At 10 MPa, diffu-sion along the free surfaces is a domi-nant diffusion path. A free surface isassociated with more open structure,and experiments showed that thejump frequency for atoms diffusingalong this defect is higher than for dif-fusion in lattice and grain boundaries(Ref. 31). With an increase in thewelding pressure to 10 MPa, effectivefree surface for contact of the matingsurfaces was increased, thus the effec-tive atomic mobility increased. Hence,DZW increases with welding pressurefrom 8 to 10 MPa. The increase in the hardness of theNi and SS with the increase in weldingpressure could be attributed to the ef-fect of work hardening. When materialis deformed, work hardening occursresulting in an increase in the hard-

ness of the material. In general, therate of work hardening is lower for hcpmetals than for cubic metals. TiA is be-ing hcp metal whereas that of Ni andSS are cubic metals. The hardness ofNi and SS were increased significantlywith the increase in the welding pres-sure from 2 to 4 MPa and above, whilethe hardness of TiA (hcp metal)doesn’t change with the welding pres-sure (Ref. 33). Along the Ni|SS inter-face, the hardness increased from Nito SS and the observed hardness was~220 HV for the DWJs processed at 2MPa, whereas it was ~330 HV for 4MPa and above welding pressures.Along the TiA|Ni diffusion zone, thehardness increased initially till itreached a maximum value anddropped toward Ni. The hardness ofNiTi intermetallic and Ni3Ti inter-metallic was ~550 and ~750 HV, re-spectively. The hardness at NiTi/Ni3Tiinterface was ~650 HV. It was also no-ticed that the fracture was expected tooccur apparently near the TiA|Ni in-terface because the hardness at theTiA|Ni interface (~750 HV) was muchhigher than the hardness at the Ni|SSinterface (~330 HV), thus the TiA|Niinterface was more brittle as comparedto the Ni|SS interface. Sam et al. (Ref.19) in their study demonstrated that

Fig. 9 — Diffusion zone width (DZW) at TiA|Ni interfaceof DWJs processed at different welding pressures.

Fig. 10 — Line sketch of the materials assembled in the DW fixture whenfurnace temperature was at A — 32C (ambient temperature); B — 750C(welding temperature).

A B


WELDING RESEARCH


DWJs failed through the TiA|Ni-alloyinterface. Here, the lateral deformation of themetals diffusion welded was of inelas-tic-type (plastic deformation); however,the plastic deformation of materials canbe correlated with the elastic propertiesof the respective materials. There arethree elastic moduli that derive fromthe spatial tri-dimensionality of theworld. The three elastic moduli repre-sent linear, planar, and volume effectson material. They are A) Young’s modu-lus (E) [resistance to changes in length];B) shear modulus (G) [resistance totransverse shift]; and C) bulk modulus(K) [resistance to compression or expan-sion]. Apart from these elastic moduli,there is another property called Pois-son’s ratio (μ) defined as the ratio of thetransverse to the axial strain. Theseelastic moduli are generally anisotropicin the case of single crystals but, sincethe engineer usually works with poly-crystalline materials in which the indi-vidual crystallites are arranged random-ly, it can generally be assumed that themechanical properties of materials areisotropic (Refs. 34–36). Althoughchange in volume is related to the bulkmodulus (K) property of the material, itcan be related to the other elastic mod-uli using the following relationships(Equations 2 and 3) (Ref. 36):

E=3K (1-2μ) (2) E=2G (1+μ) (3)

Using Equations 2 and 3, it could beperceived that Young’s modulus of ma-terials could explain the potentialcauses for significant differences inthe lateral (plastic) deformation of theTiA and SS sides of the DWJs. Theelastic (Young’s) modulus of Ti-alloy(Ti-6Al-4V) and austenitic stainlesssteels (304 SS) is ~120 GPa (Ref. 3)and ~198 GPa (Ref. 37), respectively.As Young’s modulus of Ti-6Al-4V ismuch lower as compared to Young’smodulus of 304 SS, the TiA side ofDWJs underwent an excess amount oflateral deformation as compared to theSS side of the respective DWJs; hence,it could be stated that the lateral de-formation of the material is inverselyproportional to its elastic modulus(Equation 4).

Mathematically, the slope of the later-al deformation graph (Fig. 5) was inincreasing order (Equation 5) with thewelding pressure utilized; hence, thelower-range has less slope than themiddle-range, and the middle-rangehas less slope than the higher-range ofwelding pressure.

Welding pressure lateral deformationof DWJs (5)

And it was also noticed that DWJsfractured at the TiA|Ni interface be-cause the TiA|Ni interface (~750 HV) ismuch harder than the Ni|SS interface(~330 HV). Consequently, the TiA|Niinterface is more brittle as compared tothe Ni|SS interface. In this study, it wasalso noted that the hardness differencesat the TiA|Ni interface (=750 HV–200HV=550 HV) was higher than that ofthe hardness difference at the Ni|SS in-terface (=400 HV–200 HV=200 HV).This observation testified the claimmade by Simões et al. (Ref. 38) thatDWJs fracture at the interfaces withhigh hardness differences. The fracturepath observations also supported theclaim that DWJs fractured along theTiA|Ni interface. This inference wasquite consistent with the findings of thereported literature (Refs. 17–19) as well. Base materials diffusion welded ex-perienced stresses at three successivestages. First, the base materials under-

went deformation (purely elastic innature) due to the static stressescaused by the dead load (welding pres-sure from 2 to 10 MPa) (Fig. 10A) ap-plication just before the furnace wasswitched ‘ON.’ Second, the base mate-rials underwent deformation due toincreasing dynamic thermal stresses(Fig. 10B) induced due to the linearthermal expansion of the other basematerials because of the continuousincrease in the furnace temperaturefrom room temperature to the weldingtemperature (say, 750C). Thermalstresses reach the maximum value atthe welding temperature (750C). Andthird, after reaching the designatedwelding temperature (say, 750C) thebase materials were held at the tem-perature for 60 min of the welding in-terval, i.e., during that time onward,the materials were typically subjectedto a slowly continuous deformation atconstant load (due to welding pressureapplied and the thermal stresses in-duced, see Fig. 10B) at elevated tem-peratures. This phenomenon is termedas creep (Refs. 39–41). Of these three stress fields encoun-tered by the materials diffusion welded,the effect of static stresses caused bythe dead load was trivial, while the ef-fect of creep phenomenon on the mate-rials diffusion welded was insignificantdue to the low order time scale, where-as, the effect of thermal stresses due to

Lateral deformation1

Elastic modulus(4)

Fig. 11 — Free body diagram of the base materials of DWJs when thermal stresses act atA — 32C (ambient temperature); B — 750C (welding temperature).

A B


temperature changes had a huge impacton DWJs. Although the thermal stress-es encountered by the material wereuniformly increasing from the time thefurnace was switched ‘ON’ till the timeit reached the welding temperature(750C). Thermal stress (T)TiA on TiA due tothermal expansion of SS because oftemperature change (T) is ~1867.31MPa (see Appendix), whereas the ther-mal ((stress) T)SS on SS due to ther-mal expansion of TiA is ~219.70 MPa(see Appendix). To calculate the netthermal stress (T)Net acting on thematerials, a free body diagram (Fig.11) of the materials diffusion weldedwas constructed. As both thermalstresses acting at the interface are op-posite in nature, the net thermalstress ((T)Net = (T)TiA–(T)SS=1867.31MPa–219.70 MPa) acting on the TiAand SS is ~1647.61 MPa ((T)Net).

Conclusion

This study was performed to knowthe extent of the influence of weldingpressure on the evolution of interfacialmicrostructure, mechanical properties,and fracture features of solid-state dif-fusion welded joints of Ti-6Al-4V and304 stainless steel using 200-μm-thicknickel as an interlayer. DWJs wereprocessed in vacuum at 750C (weld-ing temperature) for 60 min (weldingtime) using welding pressure in therange of 2 to 10 MPa in step of 2 MPa.Characterization of the DWJs revealedthe following: 1) Interfaces of the DWJs were freefrom the brittle Fe–Ti-based inter-metallics due to the usage of Ni inter-layer. 2) Optimal tensile strength of~560 MPa along with substantialelongation of ~12% was obtained forthe DWJs processed at 6-MPa weld-ing pressure considering the minimallateral deformation of the base met-als welded. The tensile strength wasmore or less flat for the DWJsprocessed above 6 MPa. Elongationdropped insignificantly for the DWJsprocessed above 6 MPa. 3) TiA|Ni interface (~750 HV) wasmuch harder than the Ni|SS inter-face (~330 HV). 4) It was clearly demonstrated thatthe failure of the DWJs was initiatedand propagated specifically at the

TiA|Ni interface primarily due tohigher hardness difference at theTiA|Ni interface compared to theNi|SS interface. 5) The fracture surfaces for lowerwelding pressure (2 MPa) yielded brit-tle fracture morphologies; whereas thehigher welding pressure (8 MPa) yield-ed mixed mode fractural features. 6) Apart from the welding pressureapplied, an additional stress (thermalstress) of ~1647.61 MPa (compressivein nature) was induced on the materi-als (TiA and SS) diffusion welded dueto linear thermal expansion of thematerials at the welding temperature(750C).

Considering the materials to be ho-mogeneous and with a uniform crosssection throughout the length, stress(T) in any material due to temperaturechange (T) and the restriction in thelinear thermal expansion of anothermaterial is given by the following rela-tionship:

(T)1 = (ET)2 (T)2 (T) (Refs. 42, 43)

where(T)1: Thermal stress induced in thematerial ‘1’ at temperature ‘T’

(ET)2: Young’s modulus of material ‘2’at temperature ‘T’(T)2: Coefficient of thermal expansionof the material ‘2’ at temperature ‘T’

T: Increase in the temperature

T: Welding temperature (750C)

T0: Room temperature (32C)

so, T = T–T0=750C–32C = 718C

Case 1: Thermal stress (T)TiA on TiAdue to thermal expansion of SS be-cause of temperature change (T)

(T)TiA = (ET)SS (T)SS (T)

(ET)SS = 133.37 GPa (Ref. 44)

(T)SS = 19.5 10-6 (m/m–C) (Ref. 45)

so, (T)TiA = 19.5 10-6 (m/m–C)

133.37 GPa 718C

so, (T)TiA= 1867.31 MPa (compressivein nature)

Case 2: Thermal stress (T)SS on SS dueto thermal expansion of TiA becauseof temperature change (T)

(T)SS = (ET)TA (T)TA (T)

Material properties at higher tempera-tures are less available in literature, sothe authors extrapolated using theavailable resources at hand.

(ET)TA < 30 GPa (Ref. 46)(T)TA = 10.2 10-6 (m/m–C) (Ref. 47)so, (T)SS < 10.2 × 10-6 (m/m–C ) × 30 GPa × 718C

so, (T)SS < 219.70 MPa (compressivein nature)

In all of the above calculations, ther-mal stresses induced in the nickel in-terlayer due to the expansion of bothSS and TiA were not considered for thesake of its negligible thickness (~200μm) with respect to the thickness ofSS and TiA (both ~30 mm).

1. Cobb, H. M. 2012. Dictionary of Met-als. 70, Materials Park, Ohio, ASM International. 2. Banerjee, D., and Williams, J. C.2013. Perspectives on titanium science andtechnology. Acta Materialia 61: 844–879. 3. Polmear, I. J. 2006. Light Alloys—from Traditional Alloys to Nanocrystals.299–365, 339, Oxford: Butterworth-Heinemann. 4. Yamada, M. 1996. An overview onthe development of titanium alloys fornon-aerospace application in Japan. Mate-rials Science and Engineering 213A: 8–15. 5. McGuire, M. F. 2008. Stainless Steelsfor Design Engineers. 69–90, MaterialsPark, Ohio, ASM International. 6. Rodin, M. E., Semenov, A. N., Senov,M. I., Plyshevskii, M. I., Novozhilov, S. N.,and Rassoshkina, N. S. 2013. Investigations of the strengthof steel-titanium diffusion welded joints.Welding International 27(6): 482–484. 7. He, P., Yue, X., and Zhang, J. H. 2008.Hot pressing diffusion bonding of a titani-um alloy to a stainless steel with an alu-minum alloy interlayer. Materials Science andEngineering 486A: 171–176.

WELDING RESEARCH


Appendix

References


8. Ghosh, M., Chatterjee, S., and Mishra,B. 2003. The effect of intermetallics on thestrength properties of diffusion bondsformed between Ti–5.5Al–2.4V and 304stainless steel. Materials Science and Engi-neering 363A: 268–274. 9. http://www.aerospacemetals. com/ 10. Kahraman, N., Gulenc, B., and Find-ik, F. 2007. Corrosion and mechanical-mi-crostructural aspects of dissimilar joints ofTi–6Al–4V and Al plates. International Jour-nal of Impact Engineering 34: 1423–1432. 11. Sun, Z., Annergren, I., Pan, D., andMai, T. A. 2003. Effect of laser surfaceremelting on the corrosion behaviour ofcommercially pure titanium sheet. MaterialsScience and Engineering 345A: 293–300. 12. Zeer, G. M., Zelenkova, E. G., Korole-va, Yu. P., Mikheev, A. A., and Prokopy’ev, S.V. 2013. Diffusion bonding through interlay-ers. Welding International 27(8): 638–643. 13. Hinotani, S., and Ohmori, Y. 1988.The microstructure of diffusion-bondedTi/Ni interface. Transactions of the Japan In-stitute of Metals 29(2): 116–124. 14. Kundu, S., and Chatterjee, S. 2006.Interfacial microstructure and mechanicalproperties of diffusion-bonded titanium–stainless steel joints using a nickel interlayer.Materials Science and Engineering 425A:107–113. 15. He, P., Zhang, J., Zhou, R., and Li, X.1999. Diffusion bonding technology of a ti-tanium alloy to a stainless steel web with anNi interlayer. Materials Characterization 43:287–292. 16. Kazakov, N. F. 1985. Diffusion Bond-ing of Materials. 11–12, Moscow, Mir Publishers. 17. Thirunavukarasu, G., Kundu, S.,Mishra, B., and Chatterjee, S. 2014. Effect ofbonding temperature on interfacial reactionand mechanical properties of diffusion-bonded joint between Ti-6Al-4V and 304stainless steel using nickel as an intermedi-ate material. Metallurgical and MaterialsTransactions 45A: 2067–2077. 18. Thirunavukarasu, G., Kundu, S.,Mishra, B., and Chatterjee, S. 2014. Effect ofbonding time on interfacial reaction and me-chanical properties of diffusion-bonded jointbetween Ti-6Al-4V and 304 stainless steelusing nickel as an intermediate material.Metallurgical and Materials Transactions 45A:2078–2090. 19. Sam, S., Kundu, S., and Chatterjee, S.2012. Diffusion bonding of titanium alloy to

micro-duplex stainless steel using a nickel al-loy interlayer: Interface microstructure andstrength properties. Materials and Design 40:237–244. 20. Kundu, S., Mishra, B., Olson, D. L.,and Chatterjee, S. 2013. Interfacial reactionsand strength properties of diffusion bondedjoints of Ti64 alloy and 17-4PH stainlesssteel using nickel alloy interlayer. Materialsand Design 51: 714–722. 21. Kundu, S., Sam, S., and Chatterjee, S.2013. Interfacial reactions and strengthproperties in dissimilar titanium alloy/Ni al-loy/microduplex stainless steel diffusionbonded joints. Materials Science and Engi-neering 560A: 288–295. 22. John, V. 1992. Introduction to Engi-neering Materials, p. 104, London, Macmil-lan. 23. Shanmugarajan, B., and Padmanab-ham, G. 2012. Fusion welding studies usinglaser on Ti–SS dissimilar combination, Op-tics and Lasers in Engineering 50: 1621–1627. 24. Wang, T., Zhang, B. G., and Feng, J.C. 2014. Influences of different filler metalson electron beam welding of titanium alloyto stainless steel. Transactions of NonferrousMetals Society of China 24: 108–114. 25. Balasubramanian, M. 2015. Develop-ment of processing windows for diffusionbonding of Ti−6Al−4V titanium alloy and304 stainless steel with silver as intermedi-ate layer. Transactions of Nonferrous MetalsSociety of China 25: 2932–2938. 26. Balasubramanian, M. 2016. Charac-terization of diffusion-bonded titanium al-loy and 304 stainless steel with Ag as an in-terlayer. The International Journal of Ad-vanced Manufacturing Technology 82:153–162. 27. Tomashchuk, I., Sallamand, P.,Belyavina, N., and Pilloz, M. 2013. Evolutionof microstructures and mechanical proper-ties during dissimilar electron beam weldingof titanium alloy to stainless steel via copperinterlayer. Materials Science and Engineering585A: 114–122. 28. Norouzi, E., Atapour, M., Shamanian,M., and Allafchian, A. 2016. Effect of bond-ing temperature on the microstructure andmechanical properties of Ti-6Al-4V to AISI304 transient liquid phase bonded joint. Ma-terials and Design 99: 543–551. 29. Kumar, R., and Balasubramanian, M.2015. Experimental investigation of Ti–6Al–4V titanium alloy and 304L stainless steelfriction welded with copper interlayer. De-

fence Technology 11: 65–75. 30. Fischer, T. 2009. Materials Science forEngineering Students. 97, 19–21, 112–115,San Diego, Academic Press. 31. Porter, D. A., Easterling, K. E., andSherif, M. Y. 2009. Phase Transformations inMetals and Alloys. 65, 7–9, 100–104, BocaRaton, CRC. 32. Jr Messler, R. W. 2011. The Essence ofMaterials for Engineers. 307–308, Burlington,Massachusetts: Jones & Bartlett. 33. Dieter, G. E. 1988. Mechanical Metal-lurgy. 231–233, Singapore, McGraw-Hill. 34. Kelly, P. F. 2015. Properties of Materi-als. 7–12, Boca Raton, CRC. 35. White, M. A. 2012. Physical Propertiesof Materials. 383–394, Boca Raton, CRC. 36. Lovell, M. C., Avery, A. J., and Ver-non, M. W. 1984. Physical Properties of Mate-rials. 87–92, Wokingham, Berkshire, UK,Van Nostrand Reinhold. 37. Brandes, E. A., and Brook, G. B. 1992.Smithells Metals Reference Book. 15–3, Ox-ford, Butterworth-Heinemann. 38. Simões, S., Viana, F., Koçak, M.,Ramos, A. S., Vieira, M. T., and Vieira, M. F.2011. Diffusion bonding of TiAl using reac-tive Ni/Al nanolayers and Ti and Ni foils.Materials Chemistry and Physics 128:202–207. 39. Gottstein, G. 2007. Physical Founda-tions of Materials Science. 281–282, New Del-hi, Springer. 40. Finnie, I., and Heller, W. R. 1959.Creep of Engineering Materials. 1 6, New York,McGraw-Hill. 41. Hetnarski, R. B., and Eslami, M. R.Thermal Stresses—Advanced Theory and Ap-plications. 499–510, Springer. 42. Beer, F. P., Jr Johnston, E. R., and De-Wolf, J. T. 2006. Mechanics of Materials. 74–75, New Delhi, Tata McGraw-Hill. 43. Srinath, L. S., Desayi, P., Murthy, N.S., and Ramu, S. A. 2001. Strength of Materi-als. 60–61,New Delhi, Macmillan India. 44. http://bssa.org.uk/topics.php?arti-cle=139. 45. http://asm.matweb.com/search/Speci-ficMaterial.asp?bassnum=MQ304A 46. Veiga, C., Davim, J. P., and Loureiro,A. J. R. 2012. Properties and applications oftitanium alloys: A brief review. Reviews onAdvanced Materials Science 32: 133–148. 47. http://asm.matweb.com/search/Speci-ficMaterial.asp?bassnum=MTP641.

WELDING RESEARCH



influence of welding pressure on diffusion welded joints ...g. thirunavukarasu, b. murugan, s....

Documents