Joining of aluminium alloy and galvanized steel using a controlled gas metal arc process

Journal of Manufacturing Processes 27 (2017) 179–187

Contents lists available at ScienceDirect

Journal of Manufacturing Processes journal homepage: www.elsevier.com/locate/manpro

Joining of aluminium alloy and galvanized steel using a controlled gas metal arc process A. Das a , M. Shome b , Sven-F. Goecke c , A. De a,∗ a b c

Indian Institute of Technology Bombay, Mumbai, India Tata Steel Limited, Jamshedpur, India Technical University of Applied Science, Brandenburg, Germany

a r t i c l e

i n f o

Article history: Received 14 December 2016 Received in revised form 5 April 2017 Accepted 14 April 2017 Keywords: Mixed material joint Aluminium alloy Galvanized steel Gas metal arc joining Lap joint

a b s t r a c t A precise control of the heat input is needed to restrict the development of Fe-Al brittle intermetallic compounds along the joint interface and the growth of the interface layer in mixed-material joints of aluminium and galvanized steel sheets. A novel study is presented here on the joining of aluminium alloys and galvanized steel sheets using a gas metal arc process with short-circuiting metal transfer and fast responsive control of current and voltage pulses. The influence of the processing conditions on the heat input, phase layer thickness, type of intermetallic compounds and joint strength is studied extensively. The permissible heat input is found to be in the range of 36–106 J mm−1 that has led to 0.68–6.10 ␮m thick phase layers along the joint interface with the maximum joint strength of 208 MPa. © 2017 Published by Elsevier Ltd on behalf of The Society of Manufacturing Engineers.

1. Introduction Joining of steel and aluminium alloys remains ever challenging due to their widely different thermophysical properties such as melting temperature and thermal conductivity [1,2]. The poor solubility of iron in aluminium also results in the formation and growth of Fe-Al brittle intermetallic compounds (IMC) such as Fe2 Al5 and FeAl3 along the joint interface [3]. Recent studies have showed that the thickness of the Fe-Al intermetallic phase layer can be restricted by a careful control of the heat input [4]. Several processes such as laser beam [5] and laser-arc hybrid [4] joining, gas tungsten arc [6] and gas metal arc [7] based techniques and friction stir welding [8] are examined for joining of aluminium and steel sheets. Joining processes based on pulsed current gas metal arc (GMA) are found fairly superior as they have provided fast responsive control of current pulses, metal transfer and heat input and easy adaptability to complex joint geometries and automation [7,9]. The purpose of the present study is therefore a systematic investigation to examine the suitability of an advanced pulsed current GMA process for the joining of hot-dip galvanized steel and automotive aluminium alloy sheets.

∗ Corresponding author. E-mail addresses: [email protected] (A. Das), [email protected] (M. Shome), [email protected] (S.-F. Goecke), [email protected] (A. De).

The reported studies on joining of aluminium alloys to steel sheets using GMA processes have shown a wide range of heat input to restrict the growth of the interface with Fe-Al intermetallic phases and improve the joint strength. For example, Zhang et al. [10,11] have used a heat input range of 55–91 J mm−1 to join aluminium alloy and hot-dip galvanized (GI) steel sheets, both of 1 mm in thickness, by AA4043 (Al-5%Si) filler wire. They have found the intermetallic phase layer thicknesses around 7–40 ␮m with the maximum joint strength of 96 MPa. In contrast, Zhang and Liu [12] have joined similar alloy combinations and achieved the maximum joint strength of 194 MPa using heat inputs in the range of 63–99 J mm−1 . These authors have reported the maximum intermetallic phase layer thickness of 10 ␮m [12]. Likewise, Su et al. [13] have reported the maximum joint strength of 200 MPa at a moderate heat input of 111 J mm−1 and a permissible range of phase layer thicknesses from 1 to 7 ␮m. Kang and Kim [14] have observed a similar phase layer thickness of 5 ␮m for a heat input of 112 J mm−1 in joining of aluminium alloy and GI steel sheets in the thickness ranges of 1–2 mm using Al-Si based filler wire. Yagati et al. [9] have obtained the best joint strength for a heat input of 62.75 J mm−1 in joints of 2 mm thick AA6061 and 1.2 mm thick GI steel sheets. These authors have reported the phase layer thicknesses in the range of 1.5–4.0 ␮m [9]. In joining of 1.5 mm AA5052 and 1.2 mm galvannealed (GA) steel sheets, Das et al. [7,15] have reported the highest joint strength of 73 MPa with a heat input range of 84–126 J mm−1 . The phase layer thicknesses are restricted within 1.3–2.0 ␮m [7]. Using a higher heat input range of 170–255 J mm−1 for joining of

http://dx.doi.org/10.1016/j.jmapro.2017.04.006 1526-6125/© 2017 Published by Elsevier Ltd on behalf of The Society of Manufacturing Engineers.

180

A. Das et al. / Journal of Manufacturing Processes 27 (2017) 179–187

thicker aluminium and steel sheets up to 2 mm, Murakami et al. [16] have attained the joint strength of 80 MPa and the phase layer thicknesses in the range of 0.9–2.5 ␮m. Cao et al. [2] have used a typical heat input of 200 J mm−1 for joining of thinner sheets of aluminium alloy and GI steel, and reported the phase layer thicknesses around 3–5 ␮m with the maximum joint strength of 200 MPa. These studies show the use of a wide range of heat inputs in thermal joining of aluminium alloys and galvanized steel sheets. However, the influence of heat input on joint strength and intermetallic phase layer thickness has remained inconclusive. Aluminium and galvanized steel sheets are also joined using laser beam and laser-arc hybrid joining techniques without any remarkable improvement of the joint properties. For example, Thomy and Vollertsen [4] have employed a heat input range of 47–56 J mm−1 to join 1.15 mm thick AA6016 and 1 mm thick GI steel sheets with AA4047 (Al-12%Si) filler wire. These authors have reported intermetallic phase layer thicknesses from 4 to 12 ␮m with the maximum joint strength of 184 MPa [4]. Sierra et al. [5] have used higher heat inputs of 120–150 J mm−1 to join similar sheet combinations of 1.0–1.2 mm thickness and reported the phase layer thickness of 2 ␮m and the maximum strength of around 180 MPa. Windmann et al. [17] have used another Al-Si3Mn filler wire and higher heat input ranges of 240–300 J mm−1 and, achieved the phase layer thicknesses in the range of 2–7 ␮m with the maximum joint strength of 175 MPa. In another study, Bergmann et al. [18] have joined aluminium and steel sheets with aluminized coating and reported an intermetallic phase layer thickness around 16–34 ␮m with the maximum joint strength of 135 MPa. The influence of heat input on the types of intermetallic compounds (IMC) in joining of aluminium alloys and galvanized sheets is also examined. Zhang et al. [10,11] and Su et al. [13] have reported the formation of brittle IMCs e.g. Fe2 Al5 and FeAl3 (or, Fe4 Al13 ) for a heat input range of 55–111 J mm−1 . Das et al. [7] have observed similar IMCs in joints of aluminium alloys and GA steel sheets. Higher heat inputs of 170–250 J mm−1 have led to the formation of Al-Si-Fe based IMC e.g. Fe2 Al7.4 Si [16]. In contrast, Zhang and Liu [12] have reported the formation of a ternary IMC e.g. FeAl4.5 Si at a lower heat input range of 63–99 J mm−1 . With laser assisted joining techniques, a heat input range of 47–56 J mm−1 has resulted in the formation of Fe2 Al5 along the joint interface [4]. Sierra et al. [5] have considered a higher heat input range of 120–150 J mm−1 and reported the formation of a ternary IMC phase e.g. Fe2 Al7.4 Si. In contrast to the commonly used Al-Si based filler wire, use of a AlSi3-Mn filler wire and a higher heat input range of 240–330 J mm−1 have led to the formation of Fe2 Al8 Si and FeAl3 phases [17]. In summary, the reported experimental studies have provided an understanding of the critical issues in joining of aluminium alloy and zinc-coated steel sheets using either GMA or laser-assisted hybrid arc techniques. The formation of a continuous phase layer along the joint interface in thermal joining of mixed materials such as aluminium and steel has been investigated by several authors [2–5,10,16,17]. The Fe-Al intermetallic compounds, Fex Aly , that are commonly evolved in the phase layer are found to be brittle in nature and their brittleness increases significantly with increasing concentration of aluminium [3,4,6,13]. As a result, the key aim in all joining processes is to minimize or impede the evolution and growth of especially the brittle Fex Aly compounds by controlling the heat input [2,4,7,10,16]. The ability of the advanced pulsed current GMA techniques to contain the heat input, the resulting Fe-Al intermetallic phase layer thickness at the joint interface and the formation of IMCs are not examined systematically and exhaustively. We present here an experimental study on joining of AA5754 alloy and hot-dip galvanized (GI) steel sheets using an advanced pulsed current GMA process with short-circuiting metal transfer. The primary emphasis is a quantitative understanding of the real-time

Fig. 1. Schematic diagram of experimental set-up for joining of AA5754 and GI steel sheet using gas metal arc process.

current and voltage transients and consequent heat input during the joining process. The joint bead profiles, intermetallic phase layer thickness and types of IMCs, and joint strength are examined and correlated with the processing conditions and the heat input. 2. Experimental investigation AA5754 aluminium alloy and hot-dip galvanized (GI) steel sheets are joined using a AA4043 filler wire in lap joint configuration (Fig. 1). The thicknesses of both the steel and aluminium sheets, and the filler wire diameter are 1.0 mm. A typical GMA welding power source (EWM alpha Q551) is employed to join the sample sheets using short-circuiting metal transfer that has provided fast responsive control of current and voltage pulses. Table 1 shows the chemical composition and ultimate tensile strength (UTS) of the sheets and the filler wire. Pure argon (99.999%) at a flow rate of 15 l min−1 is used as the shielding gas. The joining is done in backhand mode with the GMA torch directed on the edge of aluminium alloy and leaned at an angle of 75◦ with the horizontal. The current and voltage waveforms are monitored using a pc interfaced data acquisition system (Graphtec make GL 900-4) at concurrent sampling rates of 100 kHz. The time-averaged current (IA ), arc power (PA ) and the pulse frequency (f) are estimated as [7]  

In tn

n=0 

IA =



(1)

, tn

n=0  

PA =

In Vn tn

n=0  

,

(2)

tn

n=0

f=

1 1 

,

(3)

tn

n=0

where In , Vn and tn refer respectively to the instantaneous values of current, voltage and time, ␶ refers to either a short-circuiting or arcing period, and ␶1 is total time duration of short-circuiting and arcing periods for a complete current cycle. The variability in the estimated values of time-averaged current (IA ), arc power (PA ) and pulse frequency (f) is examined over 20 cycles in each case. The duration of the short-circuiting and arcing periods are examined from the current and voltage gradients in the corresponding transient records for a process condition. The beginning of the short-circuiting period is considered from the instant of sharp drop

A. Das et al. / Journal of Manufacturing Processes 27 (2017) 179–187

181

Fig. 2. Experimentally measured current and voltage transients corresponding to four different wire feed rates of (a) 3.5, (b) 5.0, (c) 5.5 and (d) 6.0 m min−1 during joining of GI steel sheets and aluminium alloys. tS and tA refer to short-circuiting and arcing periods. Table 1 Chemical composition (in wt%) and ultimate tensile strength (UTS) of workpiece materials and filler wire

AA5754 AA4043

Steel

Mg

Mn

Zn

Fe

Si

Cr

Cu

Ti

Al

UTS (MPa)

2.6–3.6 0.05

0.5 0.05

0.2 0.1

0.4 0.8

0.4 4.5–6.0

0.3 –

0.1 0.3

0.15 0.2

Bal. Bal.

245 200

C

Mn

P

S

Si

Cr

Cu

Ti

Al

UTS (MPa)

0.12

0.6

0.1

0.04

0.5

–

–

0.3

–

330

Table 2 Process conditions used for joining of AA5754 alloy to GI steel sheets Data Sets

wf (m min−1 )

IAV (A)

VAV (V)

v (mm s−1 )

1, 2, 3 4, 5, 6 7, 8, 9, 10 11, 12, 13, 14 15, 16, 17 18, 19, 20

3.5 4.0 4.5 5.0 5.5 6.0

39.78 (±3.2) 45.30 (±3.9) 51.15 (±3.5) 56.18 (±3.7) 61.27 (±3.6) 66.20 (±2.4)

12.46 (±0.6) 12.12 (±0.3) 12.08 (±0.8) 12.18 (±0.6) 12.32 (±0.6) 12.83 (±0.6)

5.0, 7.5, 10.0 5.0, 7.5, 10.0, 12.5 7.5, 10.0, 12.5

points with three across the intermetallic phase layer, one on the aluminium bead and one on the steel surface. The tensile strength of the joint is measured on an INSTRON 3369 universal tensile machine following the BS EN 12797:2000 standard at a cross-head speed of 1.0 mm min−1 [19]. Three tensile specimens are tested for each process condition to examine the variability in measured joint strengths. 3. Results and discussion 3.1. Estimation of heat input

in voltage and the corresponding rise of current up to the instant of sudden jump in voltage with the steep drop of current. The arcing period is considered between the two successive short-circuiting periods. The processing conditions in Table 2 are set after several trials to achieve a smooth bead appearance under visual inspection. The joint bead profiles and the respective dimensions are characterized along the transverse cross-section after polishing and etching with Keller’s reagent. The thickness of the intermetallic phase layer is measured at ten points along the joint interface using a CamScan 3200 scanning electron microscope (SEM) and an average value is considered. The nature of the Fe-Al intermetallic compounds is examined using energy dispersive spectroscopy (EDS) and CAMECA SX-5 electron probe micro-analyser (EPMA) based point analysis with an incident beam diameter of 2.0 ␮m. The EPMA analysis at an interfacial location is carried out at five

Fig. 2 shows measured current and voltage cycles during joining of AA5754 to GI steel sheets at four different wire feed rates of 3.5, 5.0, 5.5 and 6.0 m min−1 . The current and voltage transients indicate the onset of the short-circuiting period with a fast rise of current to a peak value and a drop of voltage to around 2–4 V. At the end of the short-circuiting period, the current is forced down rapidly by the high dynamic control of the power source to reignite the arc at a very low power. A high current pulse follows next for a finite duration that denotes the beginning of the arcing period with an aim to initiate fast, reproducible and constant melting of the filler wire tip. Followed by a slope down to low currents, the arcing period ends as the filler wire tip touches the workpiece forming a liquid bridge between the filler wire and melt pool. The step-wise sloped current facilitates the smooth pinching-off the liquid bridge

182

A. Das et al. / Journal of Manufacturing Processes 27 (2017) 179–187

Fig. 3. Variations of (a) time durations, and the time-averaged (b) current and (c) arc power during short-circuiting and arcing periods, and (d) pulse frequency as function of wire feed rates.

Fig. 4. Variation of estimated arc heat input as function of wire feed rate and travel speed.

from the filler wire and its deposition along the joint line during the short-circuiting period. As the arc re-ignites at a low power, the heat input is reduced that also decreases the spatter significantly. Fig. 2(a)–(d) show that the peak current increases from 130 (±4) to around 150 (±5) A with rise in wire feed rate from 3.5 to 6.0 m min−1 . The values in the parenthesis depict the variability in measurements. Fig. 2(a)–(d) also show an increase in the frequency of the current pulses as the duration of both the arcing and shortcircuiting periods reduce with increase in wire feed rate from 3.5 to 6.0 m min−1 . Fig. 3 shows the time duration, time-averaged current and arc power during short-circuiting and arcing periods, and the pulse frequency for different wire feed rates. As the wire feed rate is increased from 3.5 to 6.0 m min−1 , the short-circuiting duration reduces gently from 3.87 (±0.18) to 2.90 (±0.25) ms while the arcing period falls significantly from 13.18 (±1.02) to 7.77 (±0.55) ms as shown in Fig. 3(a). In contrast, Fig. 3(b) shows that the

time-averaged currents during both the arcing and short-circuiting periods increase from 31.87 (±2.94) to 63.83 (±2.58) A, and from 66.25 (±1.50) to 72.57 (±1.34) A, respectively. Consequently, the arc powers during both the short-circuiting and arcing periods rise from 0.18 (±0.01) to 0.30 (±0.02) kW, and from 0.50 (±0.05) to 1.05 (±0.04) kW, respectively (Fig. 3c). Fig. 3(d) shows that the pulse frequency increases from 59 (±3) to 94 (±4) Hz with increase in the wire feed rate from 3.5 to 6.0 m min−1 . Increase in pulse frequency enhances filler wire deposition rate per unit length of the joint. As a result, both the effective heat input and the deposited bead volume increase significantly at higher wire feed rates. The average heat input for a processing condition is estimated as q = (PA )/v, where PA is the time-averaged arc power considered over a complete cycle, and v and ␩ are the travel speed and the process efficiency, respectively. The value of ␩ is taken as 0.8 [20]. Fig. 4 shows a consistent rise in heat input with increase in wire feed rate and decrease in travel speed that is expected. However, the average current and arc power are significantly lower for the processing conditions adopted here due to the short-circuiting mode of metal transfer with fast responsive control of the current and voltage pulses. For example, a wire feed rate of 3.5 m min−1 has resulted in time-averaged current of around 40 A and arc power of 0.5 kW in comparison to around 76 A and 1.4 kW in a typical conventional pulsed current GMA welding processes [21,22]. As a result, the corresponding arc heat input has remained considerably lower within the selected range of processing conditions adopted in this work. 3.2. Measurement of joint bead profile Fig. 5(a–f) show the joint bead cross-section macrographs at three different wire feed rates of 4.5, 5.0 and 5.5 m min−1 , and each for two travel speeds 7.5 and 10.0 mm s−1 . The bead profiles repre-

A. Das et al. / Journal of Manufacturing Processes 27 (2017) 179–187

183

Fig. 5. Macrographs of joint bead profiles at three wire feed rates (a, d) 4.5, (b, e) 5.0 and (c, f) 5.5 m min−1 , and two travel speeds (a, b, c) 7.5 and (d, e, f) 10.0 mm s−1 . The corresponding measured values of contact angle (␤), spread length (d, mm) and bead height (h, mm) are: (a) (37.7◦ , 5.67, 1.69), (b) (33.1◦ , 5.72, 1.76), (c) (29.6◦ , 5.70, 1.85), (d) (52.1◦ , 3.67, 1.78), (e) (50.1◦ , 4.69, 1.72) and (f) (29.5◦ , 4.70, 1.73).

sent a convex shape and fairly adequate wetting of unmolten steel surface by the molten filler. The wetting is quantified further by the spread length of the filler wire deposit that is in intimate contact with the unmolten steel surface and the contact angle of the bead at the end (toe point) of the spread length. A comparison of Fig. 5(a)–(c) shows that the contact angle reduces from around 37.7◦ to 29.6◦ with rise in wire feed rate from 4.5 to 5.5 m min−1 for a constant travel speed of 7.5 mm s−1 . At a higher travel speed of 10 mm s−1 , the contact angle reduces significantly from 52.1◦ to 29.5◦ with increase in wire feed rate from 4.5 to 5.5 m min−1 [Fig. 5(d–f)]. Increase in wire feed rate at a constant travel speed results in higher heat input that improves spreading of molten filler alloy and leads to smaller contact angle. Gatzen et al. [23] has also reported contact angles of Al-Si filler deposits in the range of 24◦ to 36◦ for structurally sound joints between aluminium alloys and GI steel sheets. As the wire feed rate is increased from 4.5 to 5.5 m min−1 at a constant travel speed of 7.5 mm s−1 , the spread length of the solidified filler that is in intimate contact with the sheet surfaces remains nearly the same as shown in Fig. 5(a–c). At a higher travel speed of 10 mm s−1 , the spread length shows a gentle increase from 3.67 to 4.70 mm with rise in wire feed rate from 4.5 to 5.5 m min−1 as shown in Fig. 5(d–f). Figure 5 also shows fragments of solidified filler, unconnected to the sheets, at both ends of the spread length. Further, these unconnected fragments of solidified filler increase with decrease in travel speed and rise in wire feed rate. Yagati et al. [9] and Yang et al. [24] have suggested that boiling of zinc and its high vapour pressure can impair the intimate contact between the solidified filler and sheet surfaces especially at the root of the joint between aluminium alloy and GI steel sheets. Since increase in wire feed rate and reduction in travel speed lead to rise in heat input per unit length of joint, fragmented contacts between solidified filler and the sheets are more in Fig. 5(a–c) in comparison to that in Fig. 5(d–f). In particular, the fragmented, solidified filler at the root of the joint can form typical undercut as observed in Fig. 5(b) and (c) promoting early fracture. The influence of both wire feed rate and travel speed on the spread length of filler wire deposit on the unmolten steel surface is

Fig. 6. Measured spread length of molten filler alloy along the joint interface as function of wire feed rate for four different travel speeds of 5.0, 7.5, 10.0 and 12.5 mm s−1 .

shown in Fig. 6 for the complete range of experimental conditions. The spread length rises from 4.32 to 7.08 mm with an increase in wire feed rate from 3.5 to 5.0 m min−1 at a constant travel speed of 5.0 mm s−1· At a higher travel speed of 7.5 mm s−1 , the spread length increases with wire feed rate up to 5.0 m min−1 , and remains nearly the same with further rise in the wire feed rate as noted in Fig. 5(a–c). Fig. 6 also shows that the effect of wire feed rate on the spread length remains similar at higher values of travel speeds. Both increase in wire feed rate and decrease in travel speed enhance heat input and the deposited filler alloy per unit length of joint that contributes to higher spread length. In particular, the wire feed rate of 5.5 m min−1 and above have consistently resulted in increased mass of fragmented filler deposit at both ends of the spread length and exhibited undercut at the root of the joint. 3.3. Effect of heat input on intermetallic phase layer Fig. 7(a)–(f) show the back scatter images of the intermetallic phase layers along the joint interface for six different wire feed rates at a constant travel speed of 7.5 mm s−1 . The average phase

184

A. Das et al. / Journal of Manufacturing Processes 27 (2017) 179–187

Fig. 7. Intermetallic phase layer thickness at the AA5754 and GI steel sheet joint interface at wire feed rates of (a) 3.5, (b) 4.0, (c) 4.5, (d) 5.0, (e) 5.5 and (f) 6.0 m min−1 for a constant travel speed of 7.5 mm s−1 .

Fig. 8. Variations in (a) intermetallic phase layer thickness as function of wire feed rate and travel speed and, (b–d) elemental concentration (in at%) of Fe and Al across the interface layer, measured from unmolten steel surface, for wire feed rates of (b) 5.0, (c) 5.5, and (d) 6.0 m min−1 at a constant travel speed of 7.5 mm s−1 .

layer thicknesses have increased from 0.86 (±0.2) to 6.01 (±0.4) ␮m with rise in the wire feed rates from 3.5 to 6.0 m min−1 and resulting heat inputs from 48 (±4.3) to 90 (±3.8) J mm−1 . The intermetallic phase layer thickness shows a relatively sharp rise at the wire feed rate of 5.5 m min−1 and above. The morphology of the phase layers remains always serrated with its orientation towards the deposited bead. The serrations and sharp spikes in the phase layer morphology increase with rise in wire feed rate especially above the wire feed rate of 5.0 m min−1 . The effect of the wire feed rate on the intermetallic phase layer thickness and its morphology has remained similar for the other travel speeds.

Fig. 8(a) shows the influence of wire feed rate and travel speed on the measured intermetallic phase layer thickness for all the processing conditions. Increase in wire feed rates from 3.5 to 5.0 m min−1 at a travel speed of 5.0 mm s−1 has increased the phase layer thicknesses from 1.29 (±0.2) to 4.23 (±0.4) ␮m. The corresponding estimated heat inputs are 73 (±6.4) to 106 (±5.6) J mm−1 . With increase in the travel speed to 7.5 mm s−1 , permissible wire feed rate can be raised to 6.0 m min−1 . The phase layer thicknesses are obtained as 0.86 (±0.1) to 6.01 (±0.4) ␮m corresponding to wire feed rates of 3.5–6.0 m min−1 at a constant travel speed of 7.5 mm s−1 .

A. Das et al. / Journal of Manufacturing Processes 27 (2017) 179–187

185

Fig. 9. Variations in (a) fracture load, and (b) fracture stress for the AA5754 and GI steel sheet joints as function of wire feed rate and travel speed. (c, d) Fractured tensile specimens for wire feed rates of (c) 5.0 and (d) 5.5 m min−1 for a constant travel speed of 7.5 mm s−1 .

With further increase of the travel speed to 10 mm s−1 , the resulting heat input range is reduced as 36 (±3.2) to 68 (±2.8) J mm−1 for the wire feed rates of 3.5–6.0 m min−1 and the phase layer thicknesses are obtained from 0.52 (±0.17) to 3.73 (±0.3) ␮m. At the highest travel speed of 12.5 mm s−1 , heat inputs for the wire feed rates lower than 4.5 m min−1 are found inadequate. The phase layer thicknesses at the travel speed of 12.5 mm s−1 are obtained from 0.68 (±0.1) to 2.19 (±0.2) ␮m for the permissible range of wire feed rates of 4.5–6.0 m min−1 . The corresponding heat inputs are estimated to be around 38 (±2.3) to 54 (±2.3) J mm−1 . In summary, the intermetallic phase layer thicknesses are found in the range of 0.68–6.01 ␮m corresponding to a heat input range of 36–106 J mm−1 that is found suitable for joining the AA5754 and GI steel sheets. The elemental concentrations of Fe and Al across the phase layer thicknesses are examined further using energy dispersive X-Ray spectrometer (EDS) based point analysis for all sample joints with an aim to interpret the nature of Fe-Al intermetallic compounds. Fig. 8(b–d) show typical Fe and Al concentrations in atomic percent corresponding to wire feed rates of 5.0, 5.5 and 6.0 m min−1 and at a constant travel speed of 7.5 mm s−1 . The corresponding profiles of the intermetallic phase layers for these cases are already shown in Fig. 7. The phase layers for the wire feed rate of 4.5 m min−1 and lower have not shown any distinct Fe and Al concentration gradients, and hence, the same are not plotted. Fig. 8(b–d) show that the Al concentration near to the joint interface remains around 50.0–60.0 at% and that of Fe dwells around 35.0–40.0 at%. A comparison of Fig. 8(b–d) shows a gradual descent of the Fe concentration and consequent rise in Al concentration from the steel surface through the phase layers at higher wire feed rates and thicker layers. Diffusion of Fe atoms into the phase layers is attributed to the fairly high co-efficient of diffusion of iron in liquid aluminium [25]. A similar decaying and ascending nature of respectively Fe and Al concentrations through the intermetallic phase layers are noted for the other sample joints. An attempt is made further to characterize the concentrations of Al, Fe and also Si near to the joint interface by EPMA based point analysis as mentioned in Section 2. The search for the elemental concentration of Si is also included since the filler alloy contains Si. Table 3 shows the concentrations of those three elements at the joint interface as obtained from the EPMA based point analysis for the sample joints made at all the five wire feed rates and

Table 3 Sample EPMA results across Fe-Al interface in dissimilar joints of AA5754 alloy and hot-dip galvanized steel at a constant travel speed of 7.5 mm s−1 wf (m min−1 )

Location

Al (at%)

Fe (at%)

Si (at%)

4.0 4.5 5.0 5.5 6.0

Interface Interface Interface Interface Interface

70.55 ± 1.86 71.75 ± 1.38 73.52 ± 1.33 74.84 ± 2.51 76.94 ± 1.82

23.44 ± 3.28 20.50 ± 2.38 20.69 ± 1.56 19.47 ± 2.25 16.86 ± 1.16

6.01 ± 1.61 6.14 ± 1.25 5.79 ± 0.46 5.99 ± 0.88 6.20 ± 0.66

at a constant travel speed of 7.5 mm s−1 . Table 3 indicates a considerably high Al concentration at the interface at all the wire feed rates as also noted in Fig. 8(b–d) from EDS based analysis. However, the effect of wire feed rate on the Si concentration at the interface appears to be very little. The measured range of Al concentration at the joint interface as obtained from the EPMA based point analysis has remained nearly the same for all the processing conditions. The published literature suggests the likely formation of Fe-Al IMCs e.g. Fe2 Al5 and FeAl3 with available Al concentrations in the range of 69 ∼74 at% and 74.5–76.5 at%, respectively [26,27]. Since a more definitive phase detection analysis such as X-ray diffraction cannot be undertaken here, the presence of harmful Fe-Al IMCs along the joint interface is conceived as rational phenomena. It is also intuitive that the embrittlement effect of the Fe-Al IMCs will increase with rise in wire feed rate, resulting heat input and thicker phase layers [13]. 3.4. Influence of heat input on joint strength Fig. 9(a) and (b) show the influence of wire feed rate and travel speed on the load-bearing capacity of the joints within the range of processing conditions. The load-bearing capacity of the joints is measured in kN (Fig. 9a) that is depicted as failure load. Further, the joint strength is estimated in MPa (Fig. 9b) as the ratio of the failure load and the original joint cross-section area for each processing condition [28]. The original cross-section area resisting the load during testing is estimated as bx1 sin˛, where b is the width of the tensile specimen as shown in Fig. 10(a), and x1 and ␣ are respectively the spreading length of filler deposit and the fillet angle subtended with the spreading length as shown in Fig. 10(b) [28]. Fig. 9(a) shows an increase in failure load from 1.38 (±0.05) to 1.52 (±0.05) kN with rise in the wire feed rate from 3.5 to

186

A. Das et al. / Journal of Manufacturing Processes 27 (2017) 179–187

1 mm thick, is performed using a gas metal arc joining process with short-circuiting metal transfer. The transient current and voltage pulses are monitored to achieve the appropriate combinations of wire feed rate, travel speed and resulting heat input for reducing the growth of the Fe-Al interface layer and enhancing the joint strength. Following conclusions are made as a part of the present study.

Fig. 10. Schematic presentations of the tensile test samples of Fe-Al dissimilar joints depicting the presumed (a) bead profile and (b) spread length of filler deposit.

5.0 m min−1 at a constant travel speed of 5.0 mm s−1 . The corresponding estimated joint strength is increased by around 8% from 176.80 (±6.61) to 191.99 (±8.13) MPa as shown in Fig. 9(b). At a higher travel speed of 7.5 mm s−1 , the measured failure load increases by around 14% from 1.39 (±0.05) to 1.62 (±0.03) kN as the wire feed rate is increased from 3.5 to 5.0 m min−1 . The corresponding joint strengths are estimated as 180.99 (±5.89) to 206.51 (±3.14) MPa. In contrast, the failure load and the corresponding estimated joint strength reduce respectively to 1.36 (±0.06) kN and 172.81 (±7.9) MPa with further increase in the wire feed rate to 6.0 m min−1 . The decrease in failure load and joint strength at higher wire feed rate is attributed to increase in heat input and resulting rise in intermetallic phase layer thickness and fragmentation of filler wire deposits at the root of the joint that has acted as stress raisers [4,9]. Fig. 9(c) and (d) show two fractured tensile specimens corresponding to the wire feed rates of 5.0 and 5.5 m min−1 for a constant travel speed of 7.5 mm s−1 . The fracture locations indicate typical crack initiation from the joint root and consequent progress through the bead close to the top aluminium sheet. In dissimilar lap joints of aluminium alloys and GI steel sheets, Su et al. [13] have also reported similar fracture of the joint area under shear-tensile testing. These authors have reported intermetallic phase layer thicknesses from 4.0 to 7.0 ␮m with the maximum joint strength 188 MPa with AA 4043 (Al − 5%Si) filler wire [13]. Fig. 9(a) and (b) show further that a rise in wire feed rate from 3.5 to 5.0 m min−1 at a constant travel speed of 10 mm s−1 has enhanced the failure loads of the joints by around 10% from 1.43 (±0.05) to 1.66 (±0.05) kN and the corresponding estimated joint strength from 188.98 (±8.67) to 208.43 (±8.04) MPa. Further rise in wire feed rate to 6 m min−1 has led to a drop in the measured failure load to around 1.41 (±0.06) kN and the corresponding joint strength to 179.43 (±10.43) MPa. The nature of influence of the wire feed rate and resulting heat input on the failure load and joint strength has remained similar at all other travel speeds. The maximum joint strength is obtained as 208.43 MPa for a wire feed rate of 5.0 m min−1 and travel speed of 10 mm s−1 that is around 82% of the ultimate tensile strength of AA5754 aluminium alloy (∼253 MPa). It is noteworthy that the strengths of the joints between AA5754 and the GI steel sheets are found to be higher than that reported by the present authors in an earlier study [7] for joints between AA5952 and GA steel sheets of similar thickness ranges. 4. Conclusion A detailed experimental work on mixed material joining of automotive aluminium alloy and hot-dip galvanized steel sheets, both

(i) The selected processing conditions embodying short-circuiting metal transfer with the fast responsive current and voltage transients has allowed a significant decrease in arc power and heat input in mixed material joining of a hot-dip galvanized steel and automotive aluminium alloy sheets. (ii) A permissible heat input range of around 36–106 J mm−1 has provided suitable bead profiles with the maximum joint strength of approximately 82% of the ultimate tensile strength of the aluminium alloy. (iii) The morphology of the intermetallic phase layer is found serrated type with an orientation towards the bead side that has grown remarkably with rise in wire feed rate and heat input. (iv) Higher wire feed rate and heat input can enhance the volume of molten and vaporized zinc leading to crevice at the root and also front of the joint bead. This phenomenon is found consistent beyond a wire feed rate 5.0 m min−1 for all processing conditions selected here. This remains an oracle and further studies to examine especially the stability of metal transfer at higher wire feed rates is needed. Acknowledgments The authors acknowledge M/s EWM GmbH for providing the automotive grade alumnium alloy and hot-dip galvanized steel sheets to carry out this study. References [1] Gale WF, Totemeir TC. Smithells metals reference book. 8th ed. London: Elsevier; 2004. [2] Cao R, Yu G, Chen JH, Wang PC. Cold metal transfer joining of aluminum alloysto-galvanized mild steel. J Mater Process Technol 2013;213:1753–63. [3] Agudo L, Eyidi D, Schmaranzer CH, Arenholz E, Jank N, Bruckner J, et al. Intermetallic Fex Aly -phases in a steel/Al-alloy fusion weld. J Mater Sci 2007;42:4205–14. [4] Thomy C, Vollertsen F. Laser-MIG hybrid welding of aluminium to steel −effect of process parameters on joint properties. Weld World 2012;56:124–32. [5] Sierra G, Peyre P, Beaume FD, Stuart D, Fras G. Steel to aluminium braze welding by laser process with Al-12Si filler wire. Sci Technol Weld Join 2008;13:430–7. [6] Dong H, Hu W, Duan Y, Wang X, Dong C. Dissimilar metal joining of aluminum alloy to galvanized steel with Al-Si, Al-Cu, Al-Si-Cu and Zn-Al filler wires. J Mater Process Technol 2012;212:458–64. [7] Das A, Shome M, Das CR, Goecke SF, De A. Joining of galvannealed steel and aluminium alloy using controlled short circuiting gas metal arc welding process. Sci Technol Weld Join 2015;20:402–8. [8] Watanabe T, Takayama H, Yanagisawa A. Joining of aluminum alloy to steel by friction stir welding. J Mater Process Technol 2006;178:342–9. [9] Yagati KP, Bathe RN, Rajulapati KV, Rao KBS, Padmanabham G. Fluxless arc weld-brazing of aluminium alloy to steel. J Mater Process Technol 2014;214:2949–59. [10] Zhang HT, Feng JC, He P, Hackl H. Interfacial microstructure and mechanical properties of aluminium–zinc-coated steel joints made by a modified metal inert gas welding–brazing process. Mater Charact 2007;58:588–92. [11] Zhang HT, Feng JC, He P. Interfacial phenomena of cold metal transfer (CMT) welding of zinc coated steel and wrought aluminum. Mater Sci Technol 2008;24:1346–69. [12] Zhang H, Liu J. Microstructure characteristics and mechanical property of aluminum alloy/stainless steel lap joints fabricated by MIG welding–brazing process. Mater Sci Eng A 2011;528:6179–85. [13] Su Y, Hua X, Wu Y. Influence in alloy elements on microstructure and mechanical property of aluminum–steel lap joint made by gas metal arc welding. J Mater Process Technol 2014;214:750–5. [14] Kang M, Kim C. Joining Al 5052 alloy to aluminized steel sheet using cold metal transfer process. Mater Des 2015;81:95–103. [15] Das A, Shome M, Goecke SF, De A. Numerical modelling of gas metal arc joining of aluminium alloy and galvanized steels in lap joint configuration. Sci Technol Weld Join 2016;21:303–9.

A. Das et al. / Journal of Manufacturing Processes 27 (2017) 179–187 [16] Murakami T, Nakata K, Tong H, Ushio M. Dissimilar metal joining of aluminum to steel by MIG arc brazing using flux cored wire. ISIJ Int 2003;43:1596–602. [17] Windmann M, Rottger A, Kugler H, Theisen W, Vollertsen F. Laser beam welding of aluminum to Al-base coated high-strength steel 22MnB5. J Mater Process Technol 2015;217:88–95. [18] Bergmann JP, Stambke M, Schmidt S. Influence of aluminium coating and diffusion affecting additive on dissimilar laser joining of steel and aluminium. Phys Procedia 2013;41:190–8. [19] BS EN 12797:2000. Brazing − destructive tests of brazed joints. London: British Standard; 2004. p. 15–6. [20] Pepe N, Egerland S, Colegrove PA, Yapp D, Leonhartsberger A, Scotti A. Measuring the process efficiency of controlled gas metal arc welding processes. Sci Technol Weld Join 2011;16:412–7. [21] Pickin CG, Young K. Evaluation of cold metal transfer (CMT) process for welding aluminium alloy. Sci Technol Weld Join 2006;11:583–5. [22] Rajasekaran S, Kulkarni SD, Mallya UD, Chaturvedi RC. Droplet detachment and plate fusion characteristics in pulsed current gas metal arc welding. Weld J 1998;77:254s–69s. [23] Gatzen M, Radel T, Thomy C, Vollertsen F. Wetting and solidification characteristics of aluminium on zinc coated steel in laser welding and brazing. J Mater Process Technol 2016;238:352–60. [24] Yang S, Zhang J, Lian J, Lei Y. Welding of aluminum alloy to zinc coated steel by cold metal transfer. Mater Des 2013;49:602–12. [25] Shao L, Shi Y, Huang JK, Wu SJ. Effect of joining parameters on microstructure of dissimilar metal joints between aluminum and galvanized steel. Mater Des 2015;66:453–8. [26] Rathod MJ, Kutsuna M. Joining of aluminum alloy 5052 and low-carbon steel by laser roll welding. Weld J 2004;83:16s–26s.

187

[27] Potesser M, Schoeberl T, Antrekowitsch H, Bruckner J. The characterization of the intermetallic Fe-Al layer of steel-aluminum welding. In: Proceedings of EPD Congress −the minerals, metals & materials society. 2006. p. 167–76. [28] Makwana P, Shome M, Goecke SF, De A. Gas metal arc brazing of galvannealed steel sheets. Sci Technol Weld Join 2016;21:600–6. Atanu Das has received B Tech degree in Mechanical Engineering and M Tech degree in Production Engineering from Government Engineering College, Kalyani, India, respectively in 2009 and 2011. He has completed PhD in Engineering from the Indian Institute of Technology Bombay, India in 2016. His main research interests are in the area of welding science and technology, mixed-material joining and process modelling. Mahadev Shome is working as the head of “Material Characterization and Joining” research group at Tata Steel Limited, Jamshedpur, India and involved for the development of novel technologies, and products and processes for the steel and automotive industries. Sven-F Goecke is working as a professor in Manufacturing and Production Engineering department at the Technical University of Applied Sciences at Brandenburg an der Havel, Germany. His main research interests are in the area of arc and laser beam processing, monitoring and automation, integration of sensor systems and real-time process simulation. Amitava De is working as a professor in Mechanical Engineering department at the Indian Institute of Technology Bombay, India. His main research interests are in the area of joining science and technology, additive manufacturing and process modeling.