Heat-affected zone liquation crack on resistance spot ...dulalsaha.weebly.com/uploads/1/0/7/0/10707433/heat-affected_zone... · Heat-affected zone liquation crack on resistance spot

M A T E R I A L S C H A R A C T E R I Z A T I O N 9 3 ( 2 0 1 4 ) 4 0 – 5 1

Ava i l ab l e on l i ne a t www.sc i enced i r ec t . com

ScienceDirectwww.e l sev i e r . com/ loca te /matcha r

Heat-affected zone liquation crack on resistance

Dulal Chandra Sahaa, InSung Changb, Yeong-Do Parka,⁎aDepartment of Advanced Materials Engineering, Dong-Eui University, 995 Eomgwangno, Busanjin-gu, Busan 614-714, South KoreabAutomotive Production Development Division, Hyundai Motor Company, South Korea

A R T I C L E D A T A

⁎ Corresponding author. Tel.: +82 51 890 2290E-mail address: [email protected] (Y.-D. Pa

http://dx.doi.org/10.1016/j.matchar.2014.03.011044-5803/© 2014 Elsevier Inc. All rights rese

A B S T R A C T

Article history:Received 8 November 2013Received in revised form11 March 2014Accepted 25 March 2014

In this study, the heat affected zone (HAZ) liquation crack and segregation behavior of theresistance spot welded twinning induced plasticity (TWIP) steel have been reported. Cracksappeared in the post-welded joints that originated at the partially melted zone (PMZ) andpropagated from the PMZ through the heat affected zone (HAZ) to the base metal (BM). Thecrack length and crack opening widths were observed increasing with heat input; and thewelding current was identified to be the most influencing parameter for crack formation.Cracks appeared at the PMZ when nugget diameter reached at 4.50 mm or above; and theliquation cracks were found to occur along two sides of the notch tip in the sheet directionrather than in the electrode direction. Cracks were backfilled with the liquid films whichhas lamellar structure and supposed to be the eutectic constituent. Co-segregation of alloyelements such as, C and Mn were detected on the liquid films by electron-probemicroanalysis (EPMA) line scanning and element map which suggests that the liquid filmwas enrich of Mn and C. The eutectic constituent was identified by analyzing the calculatedphase diagram along with thermal temperature history of finite element simulation.Preliminary experimental results showed that cracks have less/no significant effect on thestatic cross-tensile strength (CTS) and the tensile-shear strength (TSS). In addition, possibleways to avoid cracking were discussed.

© 2014 Elsevier Inc. All rights reserved.

Keywords:Resistance spot weldingTWIP steelHAZ liquation crackEutectic constituentPhase diagramSimulation

1. Introduction

In order to achieve more reliability on safety, fuel economy,impact resistance of high-strength and advanced highstrength steels have been using extensively in the automotive[1,2]. Owing to outstandingmechanical properties of the TWIPsteels, it has been considered as an excellent candidate forautomotive body structure components. TWIP steel exhibitshigh strength and exceptional plastic deformation due to theformation of deformation twinning under mechanical loading[3]. Numerous researchers have devoted their interest to

(office), +82 10 6429 1860rk).

6rved.

develop the phase diagram, microstructure, and physicaldeformationmechanisms for Fe–Mn–C steels [4–7]. In additionto microstructure, and deformation mechanisms; weldability/joinability is another matter of concern which needs to beresolved to ensure proper safety to the passengers. Resistancespot welding (RSW) is the fundamental joining technique tojoin sheets for the automotive which has been evaluated bynumerous methods like by observing the microstructure,failure mode, and mechanical properties [8,9]. Recently,Mujica et al. investigated laser welding of TWIP steels and itsdissimilar combination with TRIP steels, and TIG welding of

(mobile); fax: +82 51 890 2285.

41M A T E R I A L S C H A R A C T E R I Z A T I O N 9 3 ( 2 0 1 4 ) 4 0 – 5 1

TWIP; lots of beneficial information has been obtained [10–12].Furthermore, Saha et al. [13,14] reported the microstructure,fracture morphology, segregation, and tensile properties ofthe spot welded TWIP steels; Rajinikanth et al. [15] evaluatedthe microstructure and mechanical properties of the TWIPsteels, welded with dual-phase steels for the RSW process. Itis well known that the steel with austenitic microstructure atroom temperature has a high liquation cracking susceptibilitysuch as austenitic stainless steel [16,17]. Similarly, TWIPsteels which also has an austenitic microstructure, has beenfound to be susceptible to cracking [13]. Cracks have beenobserved mostly in the mode of solidification crack (hotcrack), and liquation crack. The liquation crack occurreddue to the presence of low-melting eutectics which makescontinuous intergranular liquation. And these liquid filmshave no strength to resist thermal stresses during solidifica-tion [18]. The eutectic phase and segregation of low-meltingpoint alloys may influence the liquation crack formation[17,19]. So, it is considered to be a matter of concernunderstanding the eutectic constituent and melting temper-ature of the eutectics to find out the probability of liquationcrack.

In this research work, the liquation crack mechanism andsegregation behavior of alloys such as Mn, and C at the crackzone have been reported. Composition of the eutectic constit-uent was investigated on the basis of calculated phase diagramand finite element simulated temperature history. Further-more, liquation crack orientations and influences of thewelding parameters like weld current, welding time, electrodeforce, and electrode geometry were investigated in details.

2. Experimental Procedure

Cold rolled TWIP steel (1.4 mm) with nominal compositions of0.6C–18Mn–1.5Al, andminor alloying elements,was used in thisstudy. The welds were carried out on pedestal type medium

Fig. 1 – Schematic representation of the resistance spotwelding electrodes was used in this study; a) dome-radius(DR) type (where d is the tip diameter), andb) hemi-spherically concaved.

frequency DC (MFDC) inverter spot welder with an 8 mm tipdiameter of Cu–Cr dome-radius (DR) type (ISO 5821:2009) [20]electrode under a constant water-cooling rate of 6 l/min. Thegeometry of the electrode is schematically presented in Fig. 1a.For metallography, samples were prepared by the followingstandard metallographic procedure. The crack morphology inthe welded parts were revealed by using both nital (4% HNO3;96% ethanol) and successive color etching of nital (3% HNO3;97% ethanol) for 30 s, followed by HCl (10%) for 5 s and asolution of Na2S2O5 (10 g) in distilled water for 10 s. Thecross-tensile test (CTT) and the tensile–shear test (TST) wereperformed on Instron universal testing machine with aconstant crosshead speed of 10 mm/min. The CTT and TSTspecimens were prepared by following the ANSI/AWS/SAE/D8.9-97 standard [21].

The phase diagrams were produced using ThermoCalc®program [22] applying thermodynamic database TCFE6 foriron based materials with selective elements of Fe, C, Mn, andAl. To avoid complexity, the phases were reduced to LIQUID,FCC_A1, BCC_A2, BCC_A2#2, CEMENTIT, M7C3, and M3C. Thethermal simulations were performed using finite elementsoftware (SORPAS®) [23]; and the thermal temperaturehistories were obtained for analysis.

3. Results and Discussion

Imperfections in the RSW are greatly dependent on spotwelding schedules, electrode misalignment, welding ma-chine, unskilled operators, etc. [18]. Various welding imper-fections were observed in the weld metal, and the HAZ in thewelded part of the TWIP steel. Among them void, cavities,porosities, and HAZ liquation crack were most common typeof welding defects found in this study.

3.1. HAZ Liquation Crack

Generally, the austenitic stainless steel welds are highlysusceptible to cracking, particularly in the case of fully austeniticmode solidified weldments [24]. As the TWIP steel contains highMn which stabilizes the austenite phase; therefore, no transfor-mation occurred during the welding process; the welded zone,and the HAZwere found fully austenitic and confirmed throughEBSD experiment as reported in the previous study [14]. Richchemistry of TWIP steel enlarged the melting/solidifying tem-perature ranges; therefore, alloys were allowed to melt andsolidify over a wide temperature range [25]. At elevatedtemperature, alloys existed at the vicinity of the nugget, maymelt and form liquid films along the grain boundaries asdelineated in Fig. 2a. Temperature ranges of this zone werebetween the solidus and liquidus temperatures; this wasreferred to the PMZ. For TWIP steel, cracks were found to occurin the HAZ adjacent to the weld nugget and propagated throughthe HAZ to the BM. Zhang et al. stated that the PMZ is the crackinitiation zone, as all alloy remains in the temperature rangingbetween solidus and liquidus during welding [18]. According toHemsworth et al., when a crack is formed and filledwith liquid itcan be considered a liquation crack [26]. Robinson and Scott [27]demonstrated the prerequisite condition to identify the liqua-tion crack is the location; and that should be theHAZ, adjacent to

Fig. 2 – a) Presence of liquid films along the grains in the PMZ, and b) appearance of the liquation cracks in the HAZ [13].

42 M A T E R I A L S C H A R A C T E R I Z A T I O N 9 3 ( 2 0 1 4 ) 4 0 – 5 1

the fusion line or reheated weld metal (in case of arc welding).Another important distinguishing feature is the presence ofliquid films in the crack zone [27].

In this investigation, the cracks were found in the HAZadjacent to the fusion line and fully filled with solidified liquidmetal from the weld nugget; the liquid films/eutectic constit-uent was also observed inside the crack. Most of the cracksappeared at the two ends of the nugget with a wide openingand become narrower as they enter into the BM (see Fig. 2b). Itwas also observed that weld current have obvious effect oncrack formation; liquation cracks mostly formed when theweld current was above 5.0 kA. As the welding currentincreased (i.e. heat input increased) the tendency of thecrack was also increased [28]. Since, the size of the nuggetincreased with the weld current, which influences highertensile stress caused by the shrinkage of the nugget [29].

Cracks were revealed along the sheet direction (two sidesof the notch tip) rather than in the electrode direction (Fig. 2b).As the solidification rate was higher in the electrode direction;therefore, after being solidified in the electrode direction (dueto faster cooling rate through the bulk electrode), theremaining molten liquid was squeezed close to the fayinginterface because of the high electrode force [18]. Also coolingrate is much slower (steeper temperature gradient) throughthe sheet (horizontal direction); therefore, this uneven solid-ification rate and unbalance of the electrode pressure and thenugget pressure may initiate crack at FZ/PMZ interface andsubsequently propagates through the HAZ and eventuallythrough the BM. The PMZ of the HAZ was wetted along thegrain boundaries by lowmelting liquid film (Fig. 2a) and at thesame time the HAZ experiences high tensile strength due tosteeper temperature gradient. But these low melting liquidfilm have no ability to oppose stress in the HAZ; thereforecrack appeared. Cracks found to be filled up partially or fullyby liquid metals (mostly lower melting point alloys, andeutectics) leave an outline rather than an open space (Fig. 3).Zhang et al. demonstrated that cracks that occurred duringthe heating process may easily fill up with the liquid [18].Cracks were originated from the liquid nugget and HAZinterface (PMZ), so the liquid metal from the nugget andliquated grain boundary film may fill up crack zone. Butliquids were not sufficient to fill the crack due to their limitedamount.

Fig. 3 shows full view of the liquation crack with thepresence of the lamellar structure of low melting eutectic

constituent/divorced secondary phase. It can be seen that thevolume of the eutectics was higher at the end region of thecrack; on the other hand, eutectics were nearly absent at theroot of the crack. Near the liquation crack tip at points “a”, “b”and “c”, the crack was too narrow for weld pool liquid to flowin and backfill it. Thus, the grain-boundary liquid there wasnot affected by backfilling and solidified as eutectic asexpected. However, the liquid from the weld pool backfilledthe wide open crack root at point “d” and resulted in asolidification structure distinctly different from the eutecticnear the crack tip.

To identify and estimate the composition of the liquidfilms/eutectics in the crack area, EPMA element mapping(Fig. 4) and line scan analysis (Fig. 5) were performed along theliquation crack zone. The major alloying elements such as C,Mn and Ti were found to follow the crack zone as illustrated inFig. 4. Conversely, distribution of Al behaves oppositely in thisarea. The line scanning result shows the significant segrega-tion of C and Mn; it may form secondary phases/eutecticswhich are rich in Mn and C. Owing to strong affinity of Mn tocombine with C; the eutectics might composed of Mn and C[30]. To confirm the structure of the liquid film that existed inthe crack zone, calculated phase diagram was produced withtemperature distribution versus C percentage. Fig. 6 illustratesthat the eutectic composition (point E) point was about 5.40%of C and at a temperature of about 1120 °C. In another study,as reported by Yoo et al., for Fe–18Mn–0.6C phase diagram, theeutectic composition was found at about 4.29% C [31]; whichshifted to 5.4% C with 1.5% addition of Al. The BM contains0.6% of C; during the welding process, the temperatureelevated to above liquidus line and finally cooled at roomtemperature. Therefore, liquid films are present in the crackzone followed by the eutectic composition as indicated by thedotted line in Fig. 6. The liquid films healed the crack andweretransformed to divorced secondary eutectic constituent andsupposed to be M3C (where M is Fe, Mn) as per phase diagrampresented in Fig. 6. And the measured maximum C and Mnpercentages in liquid films were about 7.14 and 32.38,respectively; similar results were obtained for gas tungstenarc welding (GTAW) of TWIP steels [31].

3.2. Composition of the Eutectics/Liquid Films

The calculated temperature distribution from the finiteelement simulation and their corresponding microstructures

Fig. 3 –Magnified view of the liquation crack and SEMmorphologies of the lamellar eutectic constituent at different locations ofthe crack zone.

43M A T E R I A L S C H A R A C T E R I Z A T I O N 9 3 ( 2 0 1 4 ) 4 0 – 5 1

are shown in Fig. 7. The temperature distribution presented inFig. 7 shows the point wise peak temperature over the wholeprocess. During RSW process, the microstructure has trans-formed from the BM microstructure to the FZ microstructure,followed by the welding thermal profile. The studied TWIPsteel consisted of fine grained (5–7 μm) structure of fullyaustenite phase (γ) which was heated up to the meltingtemperature. The γ phase was transformed from γ to γ + M3C(where, M is Fe, Mn) at about 600 °C. At this temperature M3Cpresents as a precipitate (secondary phase) along withprimary austenite phase as can be seen in the calculatedphase diagram (Fig. 7). Further increased of temperature(about 725 °C), the phase again transformed from γ + M3C toγ; which also lead to change the grain structure to fine grainheat affected zone (FGHAZ), and then to coarse grain heataffected zone (CGHAZ). As temperature increased which

approached to the solidus temperature (point 2), the liquid(L) phase formed into the γ phase and existed as amixture of γand L; and therefore, it consisted two phases and referred tothe PMZ. The temperature of the PMZ was existed betweenthe temperature ranges of liquidus and solidus (betweenpoints 1 and 2). The common phenomenon of the PMZ wasthat the low-melting point alloys melted and existed as aliquid phase and eventually formed continuous liquid filmsaround the grains. The width of the PMZ was very narrow atabout 40 μm, and it can be varied depending on the heat inputduring RSW. The temperature ranges of this zone lies between1320 and 1420 °C and increased with C percentage ascorroborated by the phase diagram.

The Fe–Mn–C equilibrium phase diagram was shown inFig. 6 which confirmed the existence of M3C (where M is Fe,Mn) as a secondary phase in austenitic stability range at

Fig. 4 – EMPAmaps for elements of C, Mn, Al, Fe, and Ti in thecrack zone.

44 M A T E R I A L S C H A R A C T E R I Z A T I O N 9 3 ( 2 0 1 4 ) 4 0 – 5 1

700 °C; this is also in well agreement with De Cooman et al. [3].To further clarify the composition of the liquid film, thetemperature history from the simulation also was taken intoaccount. Two points were considered in the simulated result,which were A, and B; the corresponding temperature andlocation are shown in Fig. 8. The liquation crack zone was

clearly identified in the experimented cross-sectioned joint;the predicted crack zone in the simulated nugget would be inthe vicinity of the point A. The temperature history fromsimulated results is beneficial in many perspectives to predictthemicrostructure and cooling rates [32]. In this investigation,authors attempted to use the thermal profile of the nugget,PMZ, and HAZ to correlate composition of the liquid films thatexisted in the crack zone. As aforementioned, the PMZ waspreferential zone for crack formation and the temperaturerange of the PMZ coincided with simulated nodal temperatureat point B (1394 °C, Fig. 8). High volume of the liquid film waspresent at the crack tip (zones a, b and c) as delineated byFig. 3; indicating lastly solidified phases. These liquid filmshas the composition rich in Mn and C which were solidifiedafter solidifying primary γ phase through the eutectic reaction(L to γ + M3C) at about 1120 °C (Fig. 6). Therefore, M3C waspresent in the crack as a secondary eutectic constituent withprimary γ phase as per calculated phase diagram presented inFig. 6.

3.3. Effect of Welding Parameters on Crack Formation

To evaluate the influences of welding parameters on crackformation; the sequential welds were performed with varyingweld time. The welding current (peak current) was keptconstant to 6.0 kA and weld time was varied as 5 (84 ms), 7(117 ms), 10 (167 ms), 12 (200 ms), and 14 (234 ms) cycles. Thecross sectional views of the weldment are presented in Fig. 9.The results exhibited that the weld time has a noticeableeffect on crack formation; at lower weld time (5 cy (84 ms))(Fig. 9a), the crack was slightly opened which originated to thePMZ; and the crack was found only on the one coupon and oneside of the nugget which suggested that heat generation andnugget pressure were not enough to initiate fully developedcracks on the both sides of the coupons. With increasing weldtime, the cracks were found to appear on the both couponsfilled with molten metal as delineated by Fig. 9.

Fig. 10 represents the effect of the weld current on crackformation. The welds were performedwith 4, 5, and 6 kA weldcurrents while keeping other parameters constant (weldtime = 25 cy (417 ms)). It was observed that lower weldcurrent cannot open up crack due to less heat generation aswell as narrow PMZ area. Moreover, the tensile stressdeveloped in the HAZ area was not high enough at lowerweld parameters. From the above point of discussion, it can benoted here that weld current was the most influencing weldparameters to initiate the cracks; as heat generation is directlyproportional to the square of the weld current (Q = I2Rt, whereQ, I, R, and t are the heat input, weld current, resistance, andweld time, respectively).

The nugget size was the good indicator of the crackformation; throughout the study it was observed that crackswere not always appeared in the HAZ area. Cracks were onlyrevealed when the nugget size exceeded a minimum valueand crack formation was almost unavoidable as nugget sizehas to be that value to satisfy the joint strength; this was wellagreement with Luo et al.'s observation [33]. Fig. 11 representsthe crack initiation sequences with respect to the nuggetsize (for better visualization, the crack initiation zoneswere magnified, and denoted by zones A, B, and C). It was

Fig. 5 – EPMA line scans for C and Mn along the crack zone.

45M A T E R I A L S C H A R A C T E R I Z A T I O N 9 3 ( 2 0 1 4 ) 4 0 – 5 1

noticeable that when the nugget size was 4.08 mm (Fig. 11a),there was no crack in the PMZ; cracks were not found until thenugget diameter reached about 4.50 mm. When the nuggetsize ranges between 4.50 mm and 5.00 mm, small narrowcracks with negligible liquid films were observed either onecoupon (Fig. 11b) or both coupons (Fig. 11c); and above thenugget diameter of 5.00 mm the wide cracks with substantialliquid films were always found in both sides of the nugget. It isworth noting here that weld current, weld time, and nuggetsize, all have the same effect on the formation of liquationcracks, i.e. the tendency of liquation crack increased withthese parameters. All of these factors can be rationalized byconsidering the heat input; as the heat input increasedproportionally with weld time and proportional to the squareof weld current; at the same time, nugget size directlyproportional to heat input.

The electrode contact tip geometry was another importantfactor to crack formation as reported by Zhang et al. [34]. Inthis study, high tendency of crack formation was found while

Fig. 6 – Calculated phase diagram for Fe–Mn–C–Al systemshowing C percentage as a function of temperature(considering Mn: 18% and Al: 1.5%).

using dome type electrode instead of flat electrode; and theexperiment results well matched with the results as exam-ined by Zhang et al. [34]. But it was impossible to avoid cracksby using any kind of conventional electrode geometries.Fig. 12a illustrates the crack appearances using dome-radius6.0 mm electrode geometry (Fig. 1a), and Fig. 12b demon-strates the cross-sectional view of the spot weldment by usingdissimilar electrode geometries: dome 6.0 mm at bottom(Fig. 1a), and hemi-spherically concaved 6.0 mm tip diameterwith a 4.0 mm hollow space diameter at upper electrode(Fig. 1b). It is interesting to note here that cracks were alsorevealed with this type of dissimilar electrode combination;but crack orientations were completely different as observedfor similar dome-radius type electrode (Fig. 12a). It can beascribed here that high tensile strength developed in the HAZdue to the deformation of the upper coupon into thehemi-spherical groove; and narrow cracks were found in thiszone with negligible amount of liquid films inside. The crackorientation was found to be normal to the nugget tangents(where maximum tensile stress developed); on the otherhand, cracks were not identified in the other coupon placedabove the dome-radius electrode (bottom part). Anotherimportant investigation was that crack appeared in the HAZwhen the nugget diameter was about 3.75 mm; and thisobservation disclaimed previously calculated minimum nug-get diameter (4.50 mm) required for the crack formation. Dueto asymmetric electrodes, which produced high tensilestressed zone and therefore cracks were initiated in thatzone; this was the zone where hemi-spherical electrodeapplied load directly. The high tensile stress influenced toform crack when grains were substantially liquated; butcracks were not filled up with molten metal from the nugget.It was important to point out that the solidification mode wasclearly distinguished from the weld profile made by usingsymmetric electrode configuration (Fig. 12a). The hollowspace in the hemi-spherical electrode allowed solidifyingslowly as small contact area; also due to the deformation ofthe material into the hollow electrode tip area, coolingthrough water cooled electrode was not as much as fordome-radius type electrode which directly in contact withthe test coupon. As a result, nugget pressure was not

Fig. 7 – Predicted temperature distribution from fine element simulation coupled with calculated phase diagram and theirrelative microstructural changes in the RSW (the dashed lines represent the outer boundaries of the given zones).

46 M A T E R I A L S C H A R A C T E R I Z A T I O N 9 3 ( 2 0 1 4 ) 4 0 – 5 1

developed as high as to fill up the crack zone with the moltenmetal pushed from the nugget.

3.4. Contribution of Expulsion to Crack Formation

As welding current increased more heat was being generatedin the nugget, which subsequently increased the nuggetpressure and allows more liquid metal to pour into the crackzone. Due to this phenomenon, both crack length and openingcrack width were found to be increased with welding current.The contribution of the expulsion on crack formation wasillustrated in Fig. 13; where expulsion can be identified at high

Fig. 8 – The experimented nugget cross-section and the finiteelement simulated nugget showing peak temperatures inthe different regions of the weldment (including predictedcrack zone nodal temperature).

weld current (7.5 kA). By increasing the weld current, untilexpulsion, the liquation cracks become longer and wider, butthereafter, when expulsion occur, it decreased the nuggetpressure and reduces the available liquid metal, thus theliquation cracks are no more filled by liquid weld metal. Atexpulsion, the solid zone adjacent to the molten metalexperienced high tensile stress and molten metal appliesforce toward the parallel direction of the faying interface[35]; therefore, crack that appeared in the weldment whichis considered as expulsion (Fig. 13) showed different crackorientation than non-expulsion one (Fig. 2b), the crack wastraced along parallel to the faying interface in the former case.Unlike non-expulsion condition, solidified metal was notidentified inside the crack zone; furthermore, narrow crackopening width can be observed. These implies that the crackmight be initiated due to the high nugget pressure and crackwas surely formed during the heating process; thereforehealed metal was not identified as the metal being ejectedprior healing the crack. When expulsion occurred, the moltenmetal was expelled out through the interface rather thanpushed into the crack through the mushy zone.

3.5. Crack Orientation and Location

Optical microscopic examination was used to identify crackappearance and location of cracks, and their orientations.Cracks were always opened at the PMZ and propagatedthrough the HAZ to the BM; there were no cracks in thenugget without having void, micro-void, porosities, andshrinkage cavities. And cracks were identified to follow the

Fig. 9 – Crack propagation sequences as a function of the weld time (weld current: 6.0 kA); a) at 5 cy (84 ms), b) 7 cy (117 ms),c) 10 cy (167 ms), d) 12 cy (200 ms), and e) 14 cy (234 ms).

Fig. 10 – The effect of weld current on crack formation; a) at 4 kA, b) at 5 kA, and c) at 6 kA.

47M A T E R I A L S C H A R A C T E R I Z A T I O N 9 3 ( 2 0 1 4 ) 4 0 – 5 1

grain boundaries, while keeping the overall outline in theoutward direction. To detect the crack orientations: cracklengths, opening crack widths, and the angles between themain axis of the cracks and the line tangent to the fusion lineweremeasured. Fig. 14a represents the crack dimensions with

Fig. 11 – Appearances of the cracks with respect to the nugge

respect to the weld current; where both the crack length andcrack opening width were found increased with the weldcurrent. It can be seen that the crack dimensions weresignificantly increased between the weld current of 5.5 to6.0 kA; and above 6.0 kA the crack dimensions were not

t diameters, a) at 4.08 mm, b) 4.82 mm, and c) 4.91 mm.

Fig. 12 – Effect of electrode geometries on crack formation; a) using both 6.0 mm dome electrodes, and b) using 6.0 mm domeelectrode as lower and the hemi-spherically concaved electrode as upper electrode.

48 M A T E R I A L S C H A R A C T E R I Z A T I O N 9 3 ( 2 0 1 4 ) 4 0 – 5 1

changed much. On the other hand, crack angle decreaseswith weld current and average crack angle was found about60 °C. (Fig. 14b). Fig. 14c illustrates the finite element simula-tion results of the temperature contour around the nugget.An isotherm line was drawn at 800 °C temperature whichrepresents the HAZ temperature; steeper temperature gradi-ent was observed along the sheet direction, and this is in wellagreement with Zhang et al., and Gupta and De [34,36]. Zhanget al. reported that tensile stress built up tangent to theadjacent nugget isotherm [34]. Moreover, low melting pointalloying elements can melt over a wide temperature rangewhich finally produced heavy liquid film network around thegrains. So, it can be reported that the zone with steepertemperature gradient was preferential to the crack formation.

In Fig. 14d, zones A and B were the two distinct zones, asvisualized in the HAZ liquation cracked area. Zones A to Bwere expected to have substantial intergranular liquation[37] and these zones were previously described as the PMZ.The limited liquid film was observed in these zones. Beyondthe zone B, no protrusions on the grains observed, and thezones B to C were considered as a “liquated region” as statedby Lin [37].

Fig. 13 – Crack appearance and orientation at expulsion weldcondition (I = 7.5 kA).

3.6. The Significance of the Crack on Weld Strengths

In the engineering perspective, any kind of cracks or weldingdefects is completely undesirable. In the practical application, itis almost impossible to identify defects in the resistance spotwelds by using non-destructive testing methods in the produc-tion line. Generally, destructive tests are applied to evaluate theweld qualities like tensile–shear test, cross-tensile test, and peeltest. Somekind of the cracksmaynot have significant effects onweld properties; and few of them may detrimental to the weldstrength. To evaluate the significance of the liquation crack onweld strength, thewelded sampleswere progressively fracturedusing both tensile testing methods. The results were shown inFig. 15; the cross head was stopped at different displacementsduring the tests. Fig. 15a and c illustrates the fracturedcross-section at about 75% of the failure displacement, in CTTand TST, respectively. It was clearly visible that the failure wasinitiated in the notch tip and followed by the FZ/HAZ interfaceand eventually failure occurred along the sheet thicknessdirection (Fig. 15b). Conversely, the interfacial failure occurredin theTST (Fig. 15d); where the crack zone remained unaffected.So, it can be reported that the liquation cracks formed inthe RSW process has less/no significant effect on the weldproperties.

3.7. Prevention of HAZ Liquation Crack

As discussed in Section 3.6, the liquation crack doesn't haveany significant effects on weld strength; but it may enhanceother undesirable features like void formation and expulsionas reported by Luo et al. [33]. The HAZ liquation can beprevented by controlling the chemistry of the steels; it hasbeen reported that high percentage of impurities, such as Sand P, which enlarge the freezing temperature range andhence extend the liquation crack tendency [24]. So, controllinglow impurity level reduces the freezing temperature as well ascracking tendency. The grain size is another influencingfactor; it has been stated that the possibilities of the liquationcrack increased with grain sizes. In the case of finer grains,more grain boundary area, hence less liquid per unit of grainboundary area [25,26,38,39]. Another issue showed that byintroducing the ferrite stabilizer, the liquation crack could besuppressed. It is well known that Al is the ferrite stabilizer,therefore increasing the Al content solidification mode would

Fig. 14 – Crack orientations and locations; a) crack dimensions as a function of the weld current, b) angle of the cracks withrespect to the weld current, c) fine element simulated temperature contour around the nugget at the end of heating showingdense isotherm (the isotherm line was drawn at 800 °C), and d) HAZ microstructure and the crack formation zone.

49M A T E R I A L S C H A R A C T E R I Z A T I O N 9 3 ( 2 0 1 4 ) 4 0 – 5 1

change from A type (fully austenitic) to FA type (ferrite andaustenite) which has less tendency to the liquation crack[25,27]. The presence of ferrite along the grain boundariesresists wetting the boundaries; also formation of ferrite in the

Fig. 15 – Evaluation the significance of the cracks on the weld strfractured condition in the CTT, c) at 75% of total fractured strain

boundaries prevents the grain growth, which favors to havelow concentration of the liquid film around the grains. NarrowPMZ size and less tensile stress in the HAZ are beneficial toreduce cracking tendency effectively; and for this purpose less

ength; a) at 75% of total fractured strain in the CTT [13], b) atin the TST, and d) at fractured condition in the TST.

50 M A T E R I A L S C H A R A C T E R I Z A T I O N 9 3 ( 2 0 1 4 ) 4 0 – 5 1

heat input and less restrains are commended by Dupont et al.[25].

4. Conclusion

Through the experimental investigation of the liquation crackformation phenomenon and characteristics of the cracks inthe resistance spot welded TWIP steel, the following majorfindings can be concluded here:

1. TWIP steel was highly susceptible to the liquation crackformation. Intergranular characteristics of the crack andexistence of filled liquidmetal into the crack indicating crackformation occurred during heating condition. Main alloyelements such as C, Mn, and Ti showed strong segregationand followed the crack zone. The liquid film was composedof C and Mn; and formed lamellar structure of eutecticconstituent through the eutectic reaction L to γ + M3C.

2. It was found that the cracking tendency is influenced bythe heat input; high heat input enlarged the PMZ, andincrease the nugget pressure which subsequently allows toopen up the crack. The formation of liquation cracksassociated with the nugget size which is also proportionalto the heat input; it was observed that cracks were notappeared in the HAZ until a minimum nugget diameterwas reached. In this study, under a nugget diameter of4.50 mm the joints were crack free; diameter 4.50–5.00 mmwas the crack initiation zone; and diameter above 5.00 mmprovides fully developed crack.

3. Crack dimensions (crack length and crack root width)were greatly increased with the weld current. Steepertemperature gradient in the vicinity of the notch tip (sheetdirection) was identified to be the preferential location forthe crack formation.

4. The static TST and CTT results showed less/no significanteffect of liquation crack onweld strengthwhichwas believedto be due to the fine grain of the BM (less liquid networkaround grains), small crack length, and less hardness valuesin the FZ; furthermore, lower hardness helps to propagatecrack through the nugget direction in the TST or HAZ/nuggetinterface in the CTT rather than towards crack.

R E F E R E N C E S

[1] Drew S, Greg F, Michael W, Susan CA, Rita Van D, NicholasZM, et al. Climate, health, agricultural and economic impactsof tighter vehicle-emission standards. Nat Clim Chang2011;1:59–66.

[2] Kim HJ, McMillan C, Keoleian GA, Skerlos SJ. Greenhouse gasemissions payback for lightweighted vehicles usingaluminum and high-strength steel. J Ind Ecol 2010;14:929–46.

[3] De Cooman BC, Chin KG, Kim J. High Mn TWIP steels forautomotive applications. New trends and developments inautomotive system engineering. Croatia: InTech; JanezaTrdine; 2011 101–28.

[4] Schumann V. Martensitische Umwandlung in austenitischenmangan-kohlenstoff-stählen. Neue Hütte 1972;17:605–9.

[5] Qin B, Bhadeshia HKDH. Plastic strain due to twinning inaustenitic TWIP steels. Mater Sci Technol 2008;24:969–73.

[6] Allain S, Chateau JP, Bouaziz O. A physical model of thetwinning-induced plasticity effect in a high manganeseaustenitic steel. Mater Sci Eng A 2004;387–389:143–7.

[7] Remy L, Pineau A. Twinning and strain-inducedF.C.C. → H.C.P. transformation in the Fe–Mn–Cr–C system.Mater Sci Eng 1977;28:99-07.

[8] Park SS, Choi YM, Nam DG, Kim YS, Yu JH, Park YD.Evaluation of resistance spot weld interfacial fractures intensile–shear tests of TRIP 1180 steels. J Korean Weld Join Soc2008;26:81–91.

[9] Choi CY, Ji C, Nam DG, Jang J, Kim SK, Park YD. Analysis ofmicrostructure for resistance spot welded TRIP steels usingatomic force microscope. J KoreanWeld Join Soc 2013;31:43–50.

[10] Mujica L, Weber S, Thomy C, Vollertsen F. Microstructure andmechanical properties of laser welded austenitic highmanganese steels. Sci Technol Weld Join 2009;14:517–22.

[11] Mujica L, Weber S, Pinto H, Thomy C, Vollertsen F.Microstructure and mechanical properties of laser-weldedjoints of TWIP and TRIP steels. Mater Sci Eng A2010;527:2071–8.

[12] Mujica L, Weber S, Hunold G, Theisen W. Development andcharacterization of novel corrosion-resistant TWIP steels.Steel Res Int 2011;82:26–31.

[13] Saha DC, Han S, Chin KG, Choi I, Park YD. Weldabilityevaluation and microstructure analysis ofresistance-spot-welded high-Mn steel in automotiveapplication. Steel Res Int 2012;83:352–7.

[14] Saha DC, Cho Y, Park YD. Metallographic and fracturecharacteristics of resistance spot welded TWIP steels. SciTechnol Weld Join 2013;18:711–20.

[15] Rajinikanth V, Mukherjee K, Ghosh Chowdhury S, Schiebahn A,Harms A, Bleck W. Mechanical property andmicrostructure ofresistance spot welded twinning induced plasticity-dual phasesteels joint. Sci Technol Weld Join 2013;18:485–91.

[16] Ogawa T, Tsunetomi E. Hot cracking susceptibility ofaustenitic stainless steels. Weld J 1982;61:82-s–93-s.

[17] Lippold J, Baeslack W, Varol I. Heat-affected zone liquationcracking in austenitic and duplex stainless steels. Weld J1992;71:1-s–14-s.

[18] Zhang H, Senkara J. Resistance welding: fundamentals andapplications. Taylor Francis; 2006.

[19] Huang C, Kou S. Partially melted zone in aluminumwelds — solute segregation and mechanical behavior. Weld J2001;80:9–17.

[20] Kondo M, Nagata H, Nishimura A, Kokawa H. Degradationmechanism of electrode tip during resistance spot welding ofaluminium alloy sheets. Sci Technol Weld Join2011;16:126–32.

[21] Recommended practices for test methods for evaluating theresistance spot welding behavior of automotive sheet steelmaterials. American Welding Society; 2002.

[22] Andersson J, Helander T, Höglund L, Shi P, Sundman B.Thermo-Calc & DICTRA, computational tools for materialsscience. Calphad 2002;26:273-12.

[23] SORPAS version 7.0 user manual. SWANTEC Software andEngineering ApS; 2007.

[24] Shankar V, Gill TPS, Mannan SL, Sundaresan S. Solidificationcracking in austenitic stainless steel welds. Sadhana2003;28:359–82.

[25] Dopont JN, Lippold JC, Kiser SD. Solid-solution strengthenedNi-base alloys, welding metallurgy and weldability ofnickel-base alloys. Hoboken: New Jersey: John Wiley & Sons,Inc.; 2009 118–28.

[26] Hemsworth B, Boniszewski T, Eaton NF. Classification anddefinition of high temperature welding cracks in alloys. MetConstr Br Weld J 1969;1:5–15.

[27] Robinson JL, Scott MH. Liquation cracking during the weldingof austenitic stainless steels and nickel alloys. Phil Trans RSoc Lond A 1980;295:105–17.

51M A T E R I A L S C H A R A C T E R I Z A T I O N 9 3 ( 2 0 1 4 ) 4 0 – 5 1

[28] Kou S. Welding metallurgy. Hoboken: NJ: Wiley & Sons, Inc.;2003.

[29] Lang B, Sun DQ, Xuan ZZ, Qin XF. Hot cracking of resistancespot welded magnesium alloy. ISIJ Int 2008;48:77–82.

[30] Lippold JC. Kotecki DJ. Welding metallurgy and weldability ofstainless steels; 2005.

[31] Yoo J, Kim S, Park Y, Choi J, Lee C. Characteristics of the hotductility and weldability of high manganese steels. The FirstInternational Conference on High Manganese Steels; Seoul:Korea; 2011.

[32] Chan KR. Save time and money with resistance weldingsimulation software. Weld J 2008;87:32–6.

[33] Luo H, Hao C, Zhang J, Gan Z, Chen H, Zhang H.Characteristics of resistance welding magnesium alloys AZ31and AZ91. Weld J 2011;90:249–57.

[34] Zhang H, Senkara J, Wu X. Suppressing cracking in resistancewelding AA5754 by mechanical means. J Manuf Sci Eng2002;124:79–85.

[35] Senkara J, Zhang H. Cracking in spot welding aluminum alloyAA5754. Weld J 2000;79:194-s–201-s.

[36] Gupta OP, De A. An improved numerical modeling forresistance spot welding process and its experimentalverification. J Manuf Sci Eng 1998;120:246–51.

[37] Lin WE. A methodology for quantifying heat-affected zoneliquation cracking susceptibility. Ohio State University; 1991.

[38] Puzak PP, Apblett WR, Pellini WS. Hot cracking of stainlesssteel. Weld J 1956;38:9-s–17-s.

[39] Lucas J, Jackson CE. The welded heat-affected zone in nickelbase alloy 718. Weld J 1970;02:46-s–54-s.