Comparative Investigation into the Liquid Metal Embrittlement Susceptibility during Resistance Spot Welding of Zn–Al–Mg and GI Coated Advanced High Strength Steels

Abstract

In automotive industry, it is well known that cracks which are promoted by liquid metal embrittlement (LME) can occur during the resistance spot welding (RSW) of zinc-coated advanced high-strength steels (AHSS). The coating type is supposed to be one of the factors impacting LME susceptibility. Recently, Zn–Al–Mg coating is gaining increasing focus because of its enhanced corrosion resistance compared to traditional Galvanized (GI) coating. However, there is a lack of research on assessing the influence of this coating on LME susceptibility. In this study, the LME susceptibility of Zn–Al–Mg and GI coated advanced high strength steels with similar microstructure and strength are compared by hot tensile tests and RSW. The results show that Zn–Al–Mg coated samples present a more significant ductility loss than that of GI coated ones in hot tensile test, and also there are more LME cracks with large length occur in Zn–Al–Mg coated RSW joints, indicating that Zn–Al–Mg coatings have a higher LME sensitivity. The high temperature phase evolution analysis results show that Fe–Zn intermetallic compounds formed in two coatings are different, indicating that there are lower level of Fe–Zn alloying reactions in Zn–Al–Mg coating. The inadequate Fe–Zn reactions potentially facilitate the direct contact between liquid Zn and steel substrate, leading Zn–Al–Mg coating to a higher LME susceptibility.

1. Introduction

In order to satisfy an increasing demand for decreasing vehicle fuel consumption while improving crashworthiness, Advanced High Strength Steels (AHSS), such as dual phase (DP), quenching & partitioning (Q&P), complex-phase (CP), twinning-induced plasticity (TWIP), and TRIP-assisted bainitic ferrite (TBF) steels, have been increasingly used in the Body-in-white (BIW).^1,2,3) In addition, to protect the BIW from corrosive environments, a Zn-based coating is usually used to protect the automotive AHSS.^4,5) Traditionally, A zinc coating is produced via hot-dip galvanized (GI), galvannealed (GA) and electrogalvanized (EG) process. Recently, Zn–Al–Mg coating, investigated for its enhanced corrosion resistance property, is also applied to automotive AHSS gradually.^6,7)

In car-body assembly, resistance spot welding (RSW) is the most common joining process for AHSS.^8,9,10,11) In order to achieving the attractive strength-ductility combinations, a higher number of alloying elements are needed in the AHSS rather than conventional steels, and this results in several welding problems. One of the problems during the RSW process of zinc-coated AHSS is that so-called liquid mental embrittlement (LME) cracks occurring due to the liquid zinc coatings.^{12,13,14,15,16)} Although the mechanism inducing LME is still debatable in existing studies as either Zn grain boundary diffusion or liquid grain boundary penetration leading to the LME, it is widely believed that LME is a kind of intergranular cracks generally happens at high temperatures above the melting point of coating, caused by the combined effect of liquid Zn and tensile stress.^17,18,19,20)

Since LME is a not expected phenomenon resulting from the contact between the melt coating and steel substrate, so that the effects of coating and substrate steel types on the LME susceptibility are getting widely attentions. A hot tensile test and spot welding test are two common ways to evaluate LME susceptibility. In hot tension tests, the assessment is according to the ductility loss through comparison with coated and bare steels,²¹⁾ meanwhile evaluation is based on the length and number of the cracks in spot welding test.²²⁾ Many researches have been conducted to investigate the relationship between LME susceptibility and substrate materials. It has been recognized that LME sensitivity tends to become higher with the increasing strength of the steels, and cracks are commonly observed in advanced high strength steels (AHSS) during RSW and hardly reported in traditional mild steels.^14,23,24) Ling et al.²⁴⁾ found that LME susceptibility of Q&P980 was higher than that of traditional low carbon steels, due to its lower reaction with liquid Zn, higher tension stress and more LME susceptible grain boundaries. Jung et al.²⁵⁾ also found that TWIP980 has a high LME susceptibility than mild DQ steel because of its the high thermal expansion coefficient during the thermal cycle. Besides, microstructure of the steel is a more conspicuous impacting factor on LME sensitivity instead of strength among different type of AHSSs. Bhattacharya et al.²⁶⁾ investigated the influences of different starting microstructures on the LME susceptibility for different AHSSs with a fixed chemical composition by hot tension test, showing that LME susceptibility was similar for steels with a martensitic, Q&P, or TBF microstructures, while DP steel had a relatively lower susceptibility. It is suggested that the lower LME susceptibility of the DP microstructure was contributed from the finer scale of grains and less area fraction of prior austenite boundaries. Furthermore, some reports also claimed that some chemical element added in substrate steel, such as silicon, also play a role in LME crack occurrence. Tumuluru²⁷⁾ found that increasing the silicon contents in steel could facilitate LME susceptibility since silicon affected the grain boundary characters and weakened the crack resistant of steel. Hong et al.²⁸⁾ compared the LME sensitivity of TWIP steels with three different silicon concentrations and also found that adding silicon aggravated LME, and suggested that silicon suppressed the forming of α-Fe(Zn) inhibition layer resulting in a direct contact between liquid Zn and substrate, which helped liquid Zn penetrating into the substrate grain boundary. Bhattacharya et al.²⁹⁾ came to the same conclusion as Hong et al. when compared the DH and DP steels with different silicon contents. Kalashami et al.³⁰⁾ reached the similar result to the studies above and proposed that increasing silicon content would increase the depth of decarburization layer, surface ferrite grain size, and internal oxidation, leading to high LME crack susceptibility. Therefore, it can be seen that LME severity heavily depends on substrate steel due to different strength, microstructure and alloying element.

By contrast, although the relationship between LME susceptibility and coating type is also studied, but the literatures are rarer than that about the characteristics of substrate steel. Ashiri et al.³¹⁾ claimed that GI coatings express higher susceptibility than GA and EG coatings during spot welding, because low melting point of GI coating supplies the required liquid Zn for LME. Kim et al.³²⁾ compared the LME susceptibility of GI, GA and EG coatings by hot tensile test and reached the same result. The reason was interpreted by the different Fe–Zn reaction intermetallic compounds during hot tensile tests among different coatings. However, Alvarez et al.³³⁾ claimed that no significant difference was observed between GI and GA coatings in both hot tension and spot welds tests, as LME cracks can only occur just when the interface Fe–Zn metallurgical reactions were adequately slowed down and the Fe–Zn intermetallic compounds were almost the same. Meanwhile, there are hardly studies about the LME performance of Zn–Al–Mg coating. Wherein, Ghatei-Kalashami et al.³⁴⁾ studied the effect of Zn–Al–Mg on LME cracking by hot tensile test and found that although the susceptibility was high at 700°C, LME cracks were retarded at 900°C due to the developed uniform α-Fe (Zn, Al) layer in the interface. But it is worth noting that the substrate in this study was mild steel with a low LME susceptibility as known, and also there was a lack of comparison with other coating types.

According to the recent researches mentioned above. The effects of the steel substrates are well studied, but studies about the coating types are still not sufficient, especially for Zn–Al–Mg coating. In addition, few researches using both hot tensile test and spot welding test to assess the LME susceptibility at the same time, that makes it difficult to build a direct relationship between two test results. In this study, a comparative investigation into the influences of Zn–Al–Mg and GI coatings on the LME susceptibility of a DP steel with 1000 MPa grade tensile strength was discussed by both hot tensile and spot welding test. Hot tensile tests were conducted at 450–900°C, and ductility loss of samples with two type coatings were compared. At the same time, spot weld tests in a strict welding condition with a current beyond expulsion limit and a 5° electrode misalignment were carried out, crack length and number in RSW were compared to confirm the LME behaviors during an actual spot welding process. Then, the two test results of Zn–Al–Mg and GI coatings were estimated to evaluate the LME susceptibility. In order to explain the influence of the coating type on LME behavior, difference of the phase evolution process and formed Fe–Zn intermetallic compounds between Zn–Al–Mg and GI coatings were characterized.

2. Experimental Procedure

2.1. Materials

The as-received materials in this test were Zn (97 wt.%)-Al (1.5 wt.%)-Mg (1.5 wt.%) and GI coated 1000 MPa grade tensile strength DP steels, both with a sheet thickness of 1.5 mm. The Zn–Al–Mg and GI coated steels were both produced industrially through cold rolling, continuous annealing and hot dip galvanizing. The nominal chemical compositions of two materials are almost same, as listed in Table 1. The coating thicknesses for Zn–Al–Mg and GI steels were approximate 5 and 7 μm respectively.

Table 1. The nominal compositions of the as-received steels (wt.%).

Material	C	Si	Mn	P	S	Fe
GI Steel	0.10	0.35	2.38	0.01	<0.002	Bal.
Zn–Al–Mg Steel	0.12	0.36	2.34	0.01	<0.002	Bal.

2.2. Hot Tensile Tests

Hot tensile tests were used to evaluate the influence of different coating types on LME susceptibility at different temperatures in this study. The steel sheets were cut into dog-bone shaped samples using the electric wire cutting before testing. The dimension of samples is illustrated in Fig. 1(a). In order to assess the LME severity quantitatively, for Zn–Al–Mg and GI coatings, the contrast tests between the coated and the bare samples without coating were conducted. In the case of the bare samples, the coating was completely removed by soaking into a solution of 50% hydrochloric acid and 50% water for 30 s. Before testing, all the samples were cleaned with ethanol. Hot tensile tests were carried out using a Gleeble 3800 thermomechanical simulator machine under the standard atmospheric condition. According to the previous study by DiGiovanni et al.,³⁵⁾ the heating rate, holding time at the target temperature and tensile strain rate play an important role on the evaluating result of the LME sensitivity. Slower heating rates, longer holding time and lower strain rate can result in less LME susceptibility by promoting Fe–Zn alloying reaction and generating more fraction of Fe–Zn intermetallic compounds with higher melting point, which can block liquid Zn flow into the boundary of the substrate. In order to make the hot tensile test process reflect the RSW process more reliable, the recommended parameters, a 500–5000°C/s heating rate, a maximum holding time of less than 1.5 s and a strain rate between 0.1 and 1/s, were proposed in DiGiovanni’s study by comparing multiple previous studies. The thermomechanical process in this study was shown in Fig. 1(b). The target temperatures (T.T) in this test were set as 450, 600, 700, 800 and 900°C. Although a heating rate above 500°C/s is recommended by previous study, for preventing the temperature overshooting resulting from the equipment specification and ensuring the specific target temperature, the coated and bare samples were firstly heated to the temperature of 50°C lower than each T.T with heating rate of 200°C/s, and then heated to T.T separately with heating rate of 10°C/s. After holding at the T.T for 1 s, the samples were undergoing the tensile loading with a constant strain rate of 1/s until failure. The strain rate was chosen to replicate the most severe thermomechanical conditions according to the previous study. After fracture, the samples were cooled down to room temperature in atmospheric air. During the test, the temperatures of the tensile samples were controlled and measured by thermocouples welded to the center of the tensile samples, and temperature, load, displacement datas for all the samples were recorded.

Fig. 1.

Schematic diagrams of (a) the sample and (b) thermomechanical cycle for hot tensile test.

2.3. Resistance Spot Welding (RSW) Test

The resistance spot welding test was also conducted to confirm the LME susceptibility of two kinds of coatings in this study. The as-received steel sheets were cut into 125 × 40 mm specimens for welding process. The specimens were cleaned with ethanol before welding. A stationary direct current (DC) inverter-controlled spot weld machine produced by OBARA, and Cu–Cr dome-radius-type electrodes with 6-mm-tip diameters were used for the welding. Two pieces of specimens with the same type coating were welded in a two-sheet stack-up with a full overlap. For facilitating LME crack happening, the welding was carried out under an imposed welding condition with a current beyond the expulsion limit, a 5° electrode misalignment, and a shortest setting value of holding time with 20 ms. The schematic diagram and parameters of the RSW test were shown in Fig. 2 and Table 2 respectively. Five RSW samples with Zn–Al–Mg and GI coatings were welded separately.

Fig. 2.

Dimensions of spot welding for inducing LME crack.

Table 2. Welding parameters of RSW.

Electrode Force/kN	Weld Time/ms	Current/kA	Hold Time/ms
4.5	380	9.0	20

2.4. Characterization

After hot tensile tests, half of the fracture samples were cross-sectioned at the center of the length parallel to the tensile direction for microstructural characterization. The cut samples were observed to investigate the morphology of the coated layer by scanning electron microscope (SEM) and energy dispersive spectroscopy (EDS) using a Hitachi S-3400N SEM. The remaining half parts were used to determine the intermetallic compounds on the surfaces of the coated layers by X-ray diffraction (XRD). XRD were conducted by a Bruker D8 Advance A25 diffractometer with a Cu target. The scan angle range (2θ) was from 20° to 105°, and the exposure time was 0.4 s per step of 0.03°. All the RSW samples were cross-sectioned through the center of the welds parallel to the electrode misaligned direction. Then cut RSW samples were set and polished for LME crack observation. LME cracks were confirmed by a Leica DMI5000 M optical microscope. The number and depth of the cracks were measured using an image processing software ImageJ. According to previous studies, LME cracks are usually classified into four types based on happening locations,³⁶⁾ so different crack types were counted respectively in this study. The LME susceptibilities of Zn–Al–Mg and GI coatings were assessed and compared by analyzing the hot tensile test and RSW results. In hot tensile test, LME susceptibility was evaluated based on the difference in energy absorption until fracture between the coated and bare contrast samples. For RSW test, it was quantified in terms of the numbers and depths of four types of LME cracks.

3. Results

3.1. Initial Characteristics of Steel Substrates

The characteristics of steel substrates before testing were analyzed to confirm whether there was a difference in steel substrates between two type coated steels. The SE images of as-received microstructures of Zn–Al–Mg and GI coated 1000 MPa grade DP steels are shown in Fig. 3. Both of the steels had a similar dual-phase microstructures composed of martensite (M) and ferrite (F). The mechanical properties (ultimate tensile strength (UTS), and elongation to failure (EL)) of Zn–Al–Mg and GI coated steels, obtained using quasi-static tensile tests based on GB/T 228.1, are presented in Table 3. Both steels showed a similar UTS above 1000 MPa and EL above 10%, but Zn–Al–Mg coated one presented a slightly higher UTS and smaller EL than the GI coated one. From the analysis results above, it can be seen that the microstructures and mechanical properties are almost the same between substrates with two type coatings.

Fig. 3.

The SEM images of the as-received microstructure of (a) Zn–Al–Mg and (b) GI coated steels (F-ferrite, M-martensite). (Online version in color.)

Table 3. Mechanical properties of Zn–Al–Mg and GI coated steel.

Material	Ultimate Tensile Strength (MPa)	Elongation to failure (%)
Zn–Al–Mg coated steel	1070	12
GI coated steel	1010	14

3.2. Hot Tensile Behaviors

Figure 4 shows the results of the high temperature true stress-strain curves for Zn–Al–Mg, GI coated and their reference bare samples at 450, 600, 700, 800, 900°C. The bare samples with Zn–Al–Mg coating removed exhibit the relative similar tensile behaviors compared to the ones with GI coating removed at all temperatures, proving that the mechanical behaviors of the substrates with two type coatings are almost the same. Compared with the stress-strain curves of the bare samples, the Zn–Al–Mg coated ones show the similar result at 450°C and premature fractures at 600–900°C, indicating significant ductility losses within 600–900°C. By contrast, the premature fractures could only be seen at 800 and 900°C for the GI coated ones, showing that GI coated samples present a smaller ductility loss than Zn–Al–Mg coated ones. Furthermore, the ductility losses of two coatings were quantitative assessed by calculating the relative reduction of fracture energy between coated and bare samples at each tested temperature, as shown in Fig. 5(a). Here, the fracture energy was an integrating of stress and strain before fracture, and the relative reduction of fracture energy was calculated as (Ebare−Ecoated)/Ebare × 100, where Ebare and Ecoated refer to fracture energy of bare and coated samples respectively. It can be seen that the Zn–Al–Mg coated samples show a high significant loss in ductility with 48–63% fracture energy relative reductions at 600–900°C. This result is different from the result from previous research by Ghatei-Kalashami et al.,³⁴⁾ which shows that the reduction of the fracture energy decreases from about 80% at 700°C to less than 5% at 900°C. It should be noticed that the materials in previous study is IF steel, different from the DP steel used in present study, so that different results may be owing to the differences in substrate between two researches. In contrast to the Zn–Al–Mg coated samples, GI coated ones experienced relative lower energy reductions. Especially at 600 and 700°C, the samples with GI coating suffer only 9 and 13% fracture energy reductions respectively, describing nearly no ductility losses at these two temperatures. Whereas, at 800 and 900°C the ductility losses of GI coated steels increase, the relative reductions of fracture energy are 43 and 44% at these two temperatures. The result for GI coated samples is similar to previous research,²⁹⁾ the heating rate in this previous study was set as 1000°C/s, demonstrating that the temperature process in this test is appropriate for evaluating the LME sensitivity of test material. It is commonly considered that for GI coatings, the availability of liquid Zn at substrate surface requires the temperature to be higher than the peritectic temperature at 782°C, so that it can be explained that the ductility tend to be reduced only at 800 and 900°C. The peak stresses of hot tensile test as a function of temperature in the range of 450–900°C for bare, Zn–Al–Mg, and GI coated samples are shown in Fig. 5(b). From previous study, the peak stress in high temperature conditions is one of the factors on impacting LME phenomenon.^21,26) It can be observed that at 450–900°C, peak stresses are almost at the same level among samples in all coating conditions, demonstrating that the difference of the susceptibility of LME between two Zn–Al–Mg and GI coatings is not associated with obvious the tensile stress. Therefore, from the results above, Zn–Al–Mg coated samples show a higher LME susceptibility than GI coated ones in hot tensile test, and it can be inferred that the different LME sensitivity between two type coated samples are likely greater related to the coating types.

Fig. 4.

True stress-strain curves of the bare, Zn–Al–Mg and GI coated samples during hot tensile tests at: (a) 400, (b) 600, (c) 700, (d) 800, (e) 900°C. (Online version in color.)

Fig. 5.

(a) Relative reduction of fracture energy for the Zn–Al–Mg and GI coated samples, (b) peak stresses of the bare, Zn–Al–Mg and GI coated samples in 450–900°C. (Online version in color.)

The SEM micrographs of the cross-sections of the Zn–Al–Mg and GI coated samples after fracture within the temperature range of 450–900°C are shown in Fig. 6. As shown in Figs. 6(a)–6(e), the Zn–Al–Mg coated samples demonstrate a necking phenomenon at front end of the fracture area only at 450°C, and approximate no necking near fracture area in 600–900°C conditions. Cracks could be observed in the vicinity of the fracture in all Zn–Al–Mg coated samples, which indicates that LME caused by liquid Zn occurred at all test temperatures. In addition, although LME cracks are observed, the sample at 450°C presents nearly no ductility loss in mechanical property and still fails with a ductile fracture, the suppressed ductility loss at 450°C might be result from the smaller length of the cracks. By contrast, a distinct necking phenomenon are visible in the GI specimen after fracture at 450, 600 and 700°C, revealing that these ones failed in ductile fracture. But some slight cracks also take place near fractures, it can be inferred that these cracks result in the almost 10% reduction of the fracture energies at 600 and 700°C shown in Fig. 5(a). At 800 and 900°C, GI coated samples present brittle fractures similar to Zn–Al–Mg coated ones, also some LME cracks can be observed in the micrographs, but the length of cracks is shallower than the Zn–Al–Mg coated ones. SEM cross-section results are almost consistent with the stress-strain curve results, both indicate that the Zn–Al–Mg coated samples exhibit a higher LME susceptibility compared to GI ones.

Fig. 6.

SEM micrographs of the cross-sections of fractured Zn–Al–Mg coated samples after hot tensile test at (a) 450, (b) 600, (c) 700, (d) 800 and (e) 900°C, corresponding SEM micrographs of fractured GI coated specimens after tensile test at (f) 450, (g) 600, (h) 700, (i) 800, and (j) 900°C.

3.3. Comparison of LME Susceptibility of Zn–Al–Mg and GI Coatings in RSW

Figure 7 demonstrates an example of the RSW joint and the LME crack locations, it can be seen that because of the 5° electrode misalignment, weld shoulders surrounded by blue continuous and dashed shown in Fig. 7(a) is deeper than other shoulders while LME cracks are observed at these locations, as shown in Figs. 7(b) and 7(c). Meanwhile, cracks located in the indentation of the electrode are also observable under this weld condition. In the previous studies, LME cracks were classified into four types, based on location.³⁶⁾ Here, cracks located in the electrode indentation are named as Type A cracks, cracks found on the indentation shoulder are called Type D cracks. Also, the so-called Type B cracks are around the periphery of the indentation region, and Type C interfacial cracks usually occur at the lap edge of the weld joint. In prevent study, only Type A and D cracks were observed in all RSW joints. Furthermore, in order to quantitatively evaluate the LME susceptibility of Zn–Al–Mg and GI coatings during the RSW process, the length of LME cracks, determined as the distance between the surface to the crack tip, was measured for all the welded RSW joints. After length measurement, the cracks are further classified by crack length ranges. The number of Type A and D cracks in each range was counted respectively, and the results of total number for two coating types are shown in Fig. 8. It can be seen that the length distributions of Type A and D cracks present the same trend. When the crack length is less than 25 μm, the number of cracks in GI coatings is larger than Zn–Al–Mg ones, meanwhile more cracks with length larger than 25 μm are in Zn–Al–Mg coating samples than in GI coating ones. Therefore, RSW test shows the similar results with the hot tensile test, that is Zn–Al–Mg coated steels demonstrates higher LME susceptibility than GI coated ones.

Fig. 7.

An example of the RSW joint and the LME crack locations. (Online version in color.)

Fig. 8.

Statistics of LME cracks located (a) in the electrode indentation and (b) on the shoulder for Zn–Al–Mg and GI coated RSW joints. (Online version in color.)

3.4. High-temperature Phase Evolution of Zn–Al–Mg and GI Coating

Both hot tensile test and RSW test show the similar results that Zn–Al–Mg coated steels have a higher LME susceptibility than GI coated ones. As these steels have similar substrates, it is hypothesized that the LME severity is affected by coating types. Therefore, the high-temperature phase evolution of two coatings were studied to investigate the difference between two coatings at each test temperature.

The cross-section SEM micrographs with EDS point analysis positions, XRD spectra results and EDS map of element distributions of Zn–Al–Mg and GI coatings before hot tensile test are shown in Fig. 9. The average results of several EDS points are provided in Table 4. Zn–Al–Mg coating layer is observed to be a heterogeneous structure consisted of globular Zn blocks with eutectic zones presented as a lamella structure. The EDS point analysis imply that Zn content is nearly 100% in single phase zone and 90% in eutectic zone, also 3.20–4.50% Al and 3.54–3.91% Mg can be detected in eutectic zone. EDS mapping results show the distributions of Mg and Al elements in eutectic zone, and Al element in the interface between coating and substrate, indicating the existence of Fe–Al intermetallic compounds layer there. From XRD spectra, η-Zn, MgZn2, α-Al, and α-Fe are found. Based on previous research,³⁴⁾ it is known that eutectic zone includes binary and ternary eutectic zones, and the binary eutectic are made up of η-Zn and MgZn2, while ternary eutectic consists of η-Zn, MgZn2, and α-Al, it is just consisting with the XRD results. On the other hand, GI coating presents a compact and continuous Zn structure with almost 100% Zn content. EDS mapping analysis reveals that GI coating consists of pure Zn, and a thin inhibition layer of Al in the interface between coating and Fe substrate, which might be the layer of Fe2Al5 according to previous study.³²⁾ In XRD spectra results, η-Zn and α-Fe peaks are observed, but Fe2Al5 peak doesn’t appear because of its small amount.

Fig. 9.

Microstructure analyses of the coatings (a) SEM micrograph of Zn–Al–Mg coating, (b) SEM micrograph of GI coating, (c) XRD spectra results, (d) EDS map analysis of Zn–Al–Mg coating, (e) EDS map analysis of GI coating. (Online version in color.)

Table 4. The average values of EDS point analysis test for as-received coatings.

Coating type	The average of points	Fe (wt.%)	Zn (wt.%)	Al (wt.%)	Mg (wt.%)
Zn–Al–Mg	a1–a3	1.56	98.00	–	–
	a4–a6	2.06	97.23	–	–
	b1–b2	1.53	89.75	4.50	3.54
	b3–b4	2.61	89.55	3.20	3.91
GI	c1–c3	1.44	98.20	–	–
	c4–c6	1.54	98.28	–	–

The cross-section SEM micrographs with EDS analysis positions, XRD spectra results, EDS line and map of element distributions for two type coatings after hot tensile tested at 450, 600, 700, 800 and 900°C are shown in Figs. 10, 11, 12, 13, 14, and the average EDS point analysis results at each temperature are shown in Tables 4, 5, 6, 7, 8. When hot tensile tests were carried out at 450°C, the Zn–Al–Mg coating layer trends to become almost the same homogeneous structure as GI coating with the eutectic structure zones disappeared (Figs. 10(a) 10(b)). From the EDS point analysis (Table 5), Fe content in Zn–Al–Mg coating is 4.39–6.95%, which is smaller than the 8.39–10.57% in GI coating, implying a lower degree of coating and substrate interdiffusion happened in Zn–Al–Mg coating. From the XRD spectra results (Fig. 10(c)), the peaks of η-Zn, MgZn2, and α-Fe are checked out in Zn–Al–Mg coating, while δ-Zn, Γ-Zn, and α-Fe in GI coating, which also show a lower Zn–Fe alloying reaction degree in Zn–Al–Mg coating. The EDS line and map analysis results (Figs. 10(d)–10(g)) show the element distributions in two type coatings at 450°C. From the results, there is an obvious higher Fe concentration in GI coating, especially in the lower layer close to the steel substrate. Meanwhile, an inhibition layer containing Al could be detected in Zn–Al–Mg coating layer, by contrast the inhibition layer detected in the as-received GI coating has been already disappeared at this temperature. According to the results above, at 450°C the degree of Fe–Zn alloying reaction in Zn–Al–Mg coating is slower than that in GI coating, and it is suggested that the remained inhibition layer in Zn–Al–Mg coating delayed the Zn–Fe interdiffusion leading to it.

Fig. 10.

Microstructure analyses of the coatings (a) SEM micrograph of Zn–Al–Mg coating, (b) SEM micrograph of GI coating, (c) XRD spectra results, (d) EDS line analysis of Zn–Al–Mg coating, (e) EDS line analysis of GI coating, (f) EDS map analysis of Zn–Al–Mg coating, (g) EDS map analysis of GI coating after 450°C test. (Online version in color.)

Fig. 11.

Microstructure analyses of the coatings (a) SEM micrograph of Zn–Al–Mg coating, (b) SEM micrograph of GI coating, (c) XRD spectra results, (d) EDS line analysis of Zn–Al–Mg coating, (e) EDS line analysis of GI coating, (f) EDS map analysis of Zn–Al–Mg coating, (g) EDS map analysis of GI coating after 600°C test. (Online version in color.)

Fig. 12.

Microstructure analyses of the coatings (a) SEM micrograph of Zn–Al–Mg coating, (b) SEM micrograph of GI coating, (c) XRD spectra results, (d) EDS line analysis of Zn–Al–Mg coating, (e) EDS line analysis of GI coating, (f) EDS map analysis of Zn–Al–Mg coating, (g) EDS map analysis of GI coating after 700°C test. (Online version in color.)

Fig. 13.

Microstructure analyses of the coatings (a) SEM micrograph of Zn–Al–Mg coating, (b) SEM micrograph of GI coating, (c) XRD spectra results, (d) EDS line analysis of Zn–Al–Mg coating, (e) EDS line analysis of GI coating, (f) EDS map analysis of Zn–Al–Mg coating, (g) EDS map analysis of GI coating after 800°C test. (Online version in color.)

Fig. 14.

Microstructure analyses of the coatings (a) SEM micrograph of Zn–Al–Mg coating, (b) SEM micrograph of GI coating, (c) XRD spectra results, (d) EDS line analysis of Zn–Al–Mg coating, (e) EDS line analysis of GI coating, (f) EDS map analysis of Zn–Al–Mg coating, (g) EDS map analysis of GI coating after 900°C test. (Online version in color.)

Table 5. The average values of EDS point analysis test for coatings after 450°C test.

Coating type	The average of points	Fe (wt.%)	Zn (wt.%)	Al (wt.%)	Mg (wt.%)
Zn–Al–Mg	a1–a3	4.39	95.48	–	–
	a4–a6	6.95	92.50	0.33	–
GI	b1–b3	8.39	89.74	8.39	89.74
	b4–b6	10.57	88.54	10.57	88.54

Table 6. The average values of EDS point analysis test for coatings after 600°C test.

Coating type	The average of points	Fe (wt.%)	Zn (wt.%)	Al (wt.%)	Mg (wt.%)
Zn–Al–Mg	a1–a3	10.00	87.31	1.03	0.92
	a4–a6	10.16	86.67	0.70	0.91
GI	b1–b3	13.35	85.64	–	–
	b4–b6	16.14	83.15	–	–

Table 7. The average values of EDS point analysis test for coatings after 700°C test.

Coating type	The average of points	Fe (wt.%)	Zn (wt.%)	Al (wt.%)	Mg (wt.%)
Zn–Al–Mg	a1–a3	39.66	51.68	4.65	1.08
	a4–a6	23.10	73.81	0.74	–
GI	b1–b3	17.93	81.10	–	–
	b4–b6	20.66	78.73	–	–

Table 8. The average values of EDS point analysis test for coatings after 800°C test.

Coating type	The average of points	Fe (wt.%)	Zn (wt.%)	Al (wt.%)	Mg (wt.%)
Zn–Al–Mg	a1–a3	73.27	18.86	6.47	0.10
	a4–a6	25.33	69.45	0.77	0.96
GI	b1–b3	19.04	79.97	–	–
	b4–b6	58.92	40.29	–	–

At 600°C, both coating layers also exhibit the homogeneous morphologies, the same as the ones at 450°C. It can be seen that Zn–Al–Mg coating structure becomes looser with some pores and cracks in it because of the hot tensile test. By contrast, GI coating presents a relatively continuous morphology state even though pores and cracks are also observable (Figs. 11(a) 11(b)). The EDS point analysis results (Table 6) exhibit that Zn–Al–Mg coating still has a lower Fe content of 10.0–10.16% than that of 13.35–16.14% in GI coating, implying that the degree of coating and substrate interdiffusion happened in Zn–Al–Mg coating is still lower at this temperature. XRD spectra results (Fig. 11(c)) show that the peaks of η-Zn, δ-Zn, Γ-Zn, FeAl, and α-Fe are checked out in Zn–Al–Mg coating, while Γ-Zn, Γ1-Zn, and α-Fe are observed in GI coating, also indicating that the interdiffusion process in Zn–Al–Mg coating is slower, and this is consistent with the EDS point analysis results. EDS line and map analysis (Figs. 11(d)–11(g)) also show the similar results of the lower Fe content in Zn–Al–Mg coating. At the same time it can be seen that the Al contained inhibition layers disappear and Al elements are equally distributed in both coatings.

When the coatings are heated to 700°C, the Zn–Al–Mg coating layer presents a dispersed state consisting of a mixture of light contrast parts and dark contrast parts, and the dark contrast parts trend to locate at the upper layer of the coating, while the GI coating still keeps a continuous state without segregation (Figs. 12(a) 12(b)). From the EDS point analysis results (Table 7), Fe contents in both coatings are higher than that tested at 600°C suggesting that the interdiffusion of coating and substrate is deepened by the increasing of temperature. While Fe concentration of Zn–Al–Mg coating, 39.66–23.10%, become higher than that of 17.93–20.66% in GI coating at this temperature, which is different from the results of 450 and 600°C. Zn–Al–Mg and GI coating show similar XRD results (Fig. 12(c)), Γ-Zn and Γ1-Zn are both observed in both coatings, but high peaks of FeAl phase only exsit in Zn–Al–Mg coating and ZnO are formed in GI coating. The EDS line and map analysis (Figs. 12(d)–12(g)) imply that for Zn–Al–Mg coating, Mg and Al elements segregate towards the coating surface, leading to a high concentration of these two elements in the upper layer of the coating, while lower layer mainly consists of Zn. Fe exists both in upper and lower layers, and it could be seen that the dark and light contrast parts are the Fe–Al and Fe–Zn intermetallic compounds respectively. It is worth noting that the segregation of Mg was also reported in the previous study,³⁴⁾ and it was explained that the work of adhesion between Al/Fe and Zn/Fe interface increased while Mg/Fe interface decreased during the hot tensile test, so that Mg showed a strong tendency to segregate towards the coating surface in this study. Meanwhile, in present study, Al are also observed on the surface of the coating. Based on the results above, the surface of the coating mainly consists of Fe–Al compounds FeAl. It is indicated that with the increase of temperature, the FeAl were formed by Fe–Al alloying reaction, at the same time the relative position of Fe–Al compounds were moved upward to the surface because of the penetration of liquid zinc with high-density downward into the substrate, and this is consistent with the phenomenon reported in the study by Zhao et al.³⁷⁾ At the same time, for GI coating, the element distributions still keep homogeneous like at lower temperatures. The results above reveal that unlike the result at 450 and 600°C, the interdiffusion reaction between coating and substrate of Zn–Al–Mg trends to be higher than GI coating at 700°C.

Along with the temperature rising to 800°C, Zn–Al–Mg coating mainly consists of dark contrast parts while only a small fraction light contrast parts left. In contrast, the morphology of GI coating hardly changed compared to that at 700°C, even though there are some dark contrast parts scatter in the coating (Figs. 13(a) 13(b)). From the elements content results (Table 8), the dark and light contrast parts have 73.27% and 25.33% Fe concentrations separately in Zn–Al–Mg coating. For GI coating, the Fe content keeps almost the same as that at 700°C, and in the dark contrast parts Fe concentration is near 60%. From the XRD results (Fig. 13(c)), Γ-Zn and Γ1-Zn peaks are still high for both coatings, but there is a stronger intensity of α-Fe phase presented in Zn–Al–Mg coating. EDS line and map analysis results (Figs. 13(d)–13(g)) also demonstrate the drastically different element distributions in two coatings. In Zn–Al–Mg coating, Fe element is equally distributed in the dark contrast parts of the coating with a high concentration, implying a high interdiffusion degree between Fe and Zn happened. At the same time, different from 700°C the Fe–Al intermetallic layer disappeared from the surface of the coating. It is suggested that with the further rising of test temperature, Fe–Al intermetallic compounds were completely decomposed and Al element diffused into the coating again. In GI coating, Fe distribution changes little except some dark contrast parts with high Fe content appear because of the increasing temperature. According to the previous researches, the dark contrast parts would be mainly the soild solution of Zn in α-Fe, as α-Fe (Zn) phase.^17,38) Zn–Al–Mg coating almost consist of α-Fe (Zn) at this temperature, implying that the coating itself was already consumed in Fe–Zn interdiffusion at this temperature. At the same time, although α-Fe (Zn) phases are observed, GI coating mainly consist of Γ-Zn phase based on the results, indicating a low consumption degree of coating.

As the test was carried out at 900°C, for Zn–Al–Mg coating the light contrast parts decrease and dark contrast parts increase compared to 800°C, implying that the interdiffusion process proceeded. At the same time, GI coating shows quite different microstructure from that at 800°C with almost dark contrast parts in it, showing a significant increasing of interdiffusion. Two kinds of coatings come into almost the same morphology at this temperature indicating that the degrees of interdiffusion are similar (Figs. 14(a) 14(b)). The elements content results (Table 9) demonstrate that the dark contrast parts in two coatings have a similar Fe concentration of nearly 70%, implying that both of the coating almost consist of α-Fe (Zn). XRD results (Fig. 14(c)) also show that both Zn–Al–Mg and GI coatings consist of similar phases of Γ-Zn, α-Fe (Zn) and ZnO. EDS line and map analysis results (Figs. 14(d)–14(g)) show that the element distributions trends to be the same in two coatings, it is suggested that both coatings were consumed in Fe–Zn interdiffusion at this temperature.

Table 9. The average values of EDS point analysis test for coatings after 900°C test.

Coating type	The average of points	Fe (wt.%)	Zn (wt.%)	Al (wt.%)	Mg (wt.%)
Zn–Al–Mg	a1–a3	67.12	29.83	1.49	0.00
	a4–a6	27.08	71.44	0.13	0.00
GI	b1–b3	67.26	32.74	–	–
	b4–b6	74.13	25.60	–	–

4. Discussion

As the results of this study, Zn–Al–Mg coating presents a wider ductility loss temperature range (from 600 to 900°C) than that of GI coating (from 800 to 900°C) in hot tensile test, also there are more LME cracks with large length occurs in Zn–Al–Mg coated RSW joints, indicating that the LME susceptibility of Zn–Al–Mg coated sample is higher than GI ones. Moreover, two type coatings present different phase evolution processes at high temperatures, so that it is suggested that the difference of LME susceptibility could be result from the compositions of coatings.

It is well known that the liquid Zn directly contacting with the steel substrate is one of the reasons triggering LME, but the temperature for LME occurrence is usually considered to be higher than the melting temperature of Zn based coating, which is 420°C, also in this study, the ductility loss induced by LME was first observed at 600°C. It is claimed that when GI coating melts at 420°C, the solid reaction between Fe substrate and liquid Zn takes place rapidly at the interface resulting in the various solid Fe–Zn intermetallic phases formed in sequence, a serious of intermetallic phases can be confirmed in the Fe–Zn binary ζ-Zn, δ-Zn, Γ1-Zn and Γ-Zn phases. These solid intermetallic phases could play the role as a barrier to prevent liquid Zn contact with Fe directly so that LME don’t happen at high temperature. Along with the temperature rising, these compounds would be melted above the certain temperatures. Based on the Fe–Zn phase diagram, the maximum temperatures where solid phases of η-Zn, ζ-Zn, δ-Zn, and Γ-Zn exist are 419.5, 530, 665, and 782°C. All the solid phases are completely dissolved above 782°C, leading to liquid Zn exposing to the solid Fe surface beyond this temperature, so that LME hardly avoided beyond this temperature. Then with the cooling of samples, a series of solidification process and peritectic reactions will take place, and each solid phase could be re-formed again at temperatures mentioned above. Therefore, the constituent Fe–Zn compound phases at high test temperatures could be predicted according to the phases existing in the coatings after tensile test. If the predicted Fe–Zn compounds appeared as liquid phase at some temperature, it could be suggested that liquid Zn exist and LME were probably induced at that temperature. Based on the theory above, the difference of LME susceptibility between two coatings can be explained by comparing the Fe–Zn phases after hot tensile test at each test temperature.

The results of existing and predicted phases are shown in Table 10. At 450 and 600°C, the differences of the Fe–Zn phases between two coatings could be observed. η phase, which was liquid state at these temperatures, are only observed in Zn–Al–Mg coatings, indicating that the Zn–Al–Mg coated samples exposing to liquid Zn and experiencing the risk of LME. For Zn–Al–Mg coating, ductility loss doesn’t happen at 450°C, but some LME cracks are observed in the samples after tensile test (Fig. 6(a)), implying that small number and little length of LME cracks may not be enough to cause reduction of mechanical properties. At 600°C, the ductility loss occurs with long LME cracks in the Zn–Al–Mg samples shown in Fig. 6(b). Meanwhile, for GI coating, no LME cracks or ductility losses are seen in the samples because of no liquid phases exist at these low temperatures. At 700°C, both coatings consist of liquid phases but LME-induced ductility loss is only observed in Zn–Al–Mg coated sample. For GI coating, although there are some little LME cracks take places in GI coated sample (Fig. 6(h)), but the changing of mechanical property is not obvious, and the reason is thought to be the same as that of Zn–Al–Mg coating at 450°C. When the heating temperatures reach 800 and 900°C, which are higher than the maximum temperatures where all the Fe–Zn solid phases exist, the steel substrate totally expose to liquid Zn for two coatings, and as a result dramatic ductility losses happened in both Zn–Al– Mg and GI coated samples.

Table 10. The existing and predicted phases at each test temperature for two type coatings.

Test temperature (°C)	Coating Type	Constituent phases after test	Predicted phases at test temperature
450	Zn–Al–Mg	η	Liquid (η)
	GI	δ, Γ	δ, Γ
600	Zn–Al–Mg	η, δ, Γ	Liquid (η), δ, Γ
	GI	Γ, Γ1	Γ, δ
700	Zn–Al–Mg	Γ, Γ1	Γ, Liquid (δ)
	GI	Γ, Γ1	Γ, Liquid (δ)
800	Zn–Al–Mg	Γ, Γ1, α-Fe	Liquid (δ+Γ), α-Fe
	GI	Γ, Γ1, α-Fe	Liquid (δ+Γ), α-Fe
900	Zn–Al–Mg	Γ, Γ1, α-Fe	Liquid (δ+Γ), α-Fe
	GI	Γ, Γ1, α-Fe	Liquid (δ+Γ), α-Fe

Based on the discussion above, it is clear that the differences of LME susceptibility between two coatings results from the Fe–Zn phases. Some previous researches show that the existence of Al suppresses Fe–Zn alloying reactions by forming a Fe–Al intermetallic layer to obstruct the interdiffusion between Fe and Zn.^37,38) In this study, the Fe–Al inhibition layer could be observed at 450°C, so the key theories related to this behavior can be explained and summarized as follows: Zn–Al–Mg is an Al-enriched coating layer rather than GI coating, when two type coatings are heated above the melting temperature, the coating and substrate atoms trend to interdiffuse at the interface. In Al-enriched Zn–Al–Mg coating, the Fe–Al intermetallic layer formed firstly at the coating/substrate interface, so that the process of Fe–Zn alloying reaction is retarded and the formation of solid Fe–Zn phases are suppressed. On the other hand, in GI coating, Fe–Zn interdiffusion and the formation of solid Fe–Zn intermetallic phases happened smoothly. Because of this, more liquid Zn are available to induce LME in Al-enriched Zn–Al–Mg coating.

5. Conclusions

This study compared the influences of Zn–Al–Mg and GI coatings on the LME susceptibility with the same substrate of a 1000 MPa grade DP steel. The LME susceptibility was assessed by both hot tensile and spot welding test. The main conclusions are summarized as following:

(1) In hot tensile test, Zn–Al–Mg coating presents a wider ductility loss temperature range (from 600 to 900°C) than that of GI coating (from 800 to 900°C), implying that the LME sensitivity of Zn–Al–Mg coating is higher than GI coating.

(2) Resistance spot welding test showed the similar results with the hot tensile test, that is Zn–Al–Mg coated steels demonstrates more serious LME cracks (more cracks with larger length) compared to GI coating ones.

(3) SEM and XRD analysis results indicated the different Fe–Zn intermetallic phase evolution processes between two coatings at high temperatures caused the difference of LME sensitivity between two coatings. Especially at 450 and 600°C, η phase, which was liquid state at these temperatures, could be observed in Zn–Al–Mg coatings, indicating that the Zn–Al–Mg coated samples experiencing the more risk of LME than GI coated ones.

(4) Zn–Al–Mg is an Al-enriched coating layer. It is hypothesized that in Al-riched coating, the Fe–Al intermetallic layer formed firstly at the coating/substrate interface at high temperature, so that the process of Fe–Zn alloying reaction is retarded and the formation of solid Fe–Zn phases are suppressed. The inadequate Fe–Zn reactions facilitated the direct contact between liquid Zn and steel substrate, leading Zn–Al–Mg coating to a higher LME susceptibility.

References

• 1)   J.  Schmitt and  T.  Iung: C. R. Phys., 19 (2018), 641. https://doi.org/10.1016/j.crhy.2018.11.004
• 2)   J.  Galán,  L.  Samek,  P.  Verleysen,  K.  Verbeken and  Y.  Houbaert: Rev. Metal., 48 (2012), 118. https://doi.org/10.3989/revmetalm.1158
• 3)   D.  Matlock and  J.  Speer: Mater. Manuf. Process., 25 (2010), 7. https://doi.org/10.1080/10426910903158272
• 4)   A. R.  Marder: Prog. Mater. Sci., 45 (2000), 191. https://doi.org/10.1016/S0079-6425(98)00006-1
• 5)   S. M. A.  Shibli,  B. N.  Meena and  R.  Remya: Surf. Coat. Technol., 262 (2015), 210. https://doi.org/10.1016/j.surfcoat.2014.12.054
• 6)   C.  Yao,  H.  Lv,  T.  Zhu,  W.  Zheng,  X.  Yuan and  W.  Gao: J. Alloy. Compd., 670 (2016), 239. https://doi.org/10.1016/j.jallcom.2016.02.026
• 7)   D.  Thierry,  D.  Persson,  G.  Luckeneder and  K.  Stellnberger: Corros. Sci., 148 (2019), 338. https://doi.org/10.1016/j.corsci.2018.12.033
• 8)   M.  Pouranvari and  S. P. H.  Marashi: Sci. Technol. Weld. Join., 18 (2013), 361. https://doi.org/10.1179/1362171813Y.0000000120
• 9)   I.  Chang,  Y.  Cho,  H.  Park and  D.  So: J. Weld. Join., 34 (2016), 21 (in Korean). https://doi.org/10.5781/jwj.2016.34.1.21
• 10)   N. T.  Williams and  J. D.  Parker: Int. Mater. Rev., 49 (2004), 77. https://doi.org/10.1179/095066004225010541
• 11)   A.  Al-Mukhtar: Adv. Eng. Forum, 17 (2016), 31. https://doi.org/10.4028/www.scientific.net/AEF.17.31
• 12)   S. S. S.  Guraja,  S.  Patra,  K. S.  Arora and  N.  Kumar: Trans. Indian Inst. Met., 75 (2022), 1695. https://doi.org/10.1007/s12666-022-02569-7
• 13)   Z.  Ling,  T.  Chen,  L.  Kong,  M.  Wang,  H.  Pan and  M.  Lei: Metall. Mater. Trans. A, 50 (2019), 5128. https://doi.org/10.1007/s11661-019-05388-6
• 14)   D.  Bhattacharya: Mater. Sci. Technol., 34 (2018), 1809. https://doi.org/10.1080/02670836.2018.1461595
• 15)   W.  Jeon,  A.  Sharma and  J. P.  Jung: Metals, 10 (2020), 787. https://doi.org/10.3390/met10060787
• 16)   M. H.  Razmpoosh,  C.  DiGiovanni,  Y. N.  Zhou and  E.  Biro: Prog. Mater. Sci., 121 (2021), 100798. https://doi.org/10.1016/j.pmatsci.2021.100798
• 17)   H.  Kang,  L.  Cho,  C.  Lee and  B. C.  De Cooman: Metall. Mater. Trans. A, 47 (2016), 2885. https://doi.org/10.1007/s11661-016-3475-x
• 18)   S. P.  Murugan,  J. B.  Jeon,  C.  Ji and  Y.  Park: Weld. World, 64 (2020), 1957. https://doi.org/10.1007/s40194-020-00975-3
• 19)   C.  DiGiovanni,  A.  Ghatei Kalashami,  E.  Biro and  N. Y.  Zhou: Materialia, 18 (2021), 101153. https://doi.org/10.1016/j.mtla.2021.101153
• 20)   H.  Lee,  M. C.  Jo,  S. S.  Sohn,  S.  Kim,  T.  Song,  S.  Kim,  H. S.  Kim,  N. J.  Kim and  S.  Lee: Mater. Charact., 147 (2019), 233. https://doi.org/10.1016/j.matchar.2018.11.008
• 21)   C.  Beal,  X.  Kleber,  D.  Fabregue and  M.  Bouzekri: Mater. Sci. Eng. A, 543 (2012), 76. https://doi.org/10.1016/j.msea.2012.02.049
• 22)   E.  Wintjes,  C.  DiGiovanni,  L.  He,  E.  Biro and  N. Y.  Zhou: Weld. World, 63 (2019), 807. https://doi.org/10.1007/s40194-019-00712-5
• 23)   Y. G.  Kim,  I. J.  Kim,  J. S.  Kim,  Y. I.  Chung and  D. Y.  Choi: Mater. Trans., 55 (2014), 171. https://doi.org/10.2320/matertrans.M2013244
• 24)   Z.  Ling,  M.  Wang,  L.  Kong and  K.  Chen: Mater. Des., 195 (2020), 109055. https://doi.org/10.1016/j.matdes.2020.109055
• 25)   G.  Jung,  I. S.  Woo,  D. W.  Suh and  S.  Kim: Met. Mater. Int., 22 (2016), 187. https://doi.org/10.1007/s12540-016-5579-7
• 26)   D.  Bhattacharya,  L.  Cho,  E.  van der Aa,  A.  Pichler,  N.  Pottore,  H.  Ghassemi-Armaki,  K. O.  Findley and  J. G.  Speer: Mater. Sci. Eng. A, 804 (2021), 140391. https://doi.org/10.1016/j.msea.2020.140391
• 27)   M.  Tumuluru: Weld. J., 98 (2019), 351-s. https://doi.org/10.29391/2019.98.029
• 28)   S.  Hong,  J.  Kang,  D.  Kim and  S.  Kim: Surf. Coat. Technol., 393 (2020), 125809. https://doi.org/10.1016/j.surfcoat.2020.125809
• 29)   D.  Bhattacharya,  L.  Cho,  D.  Marshall,  M.  Walker,  E.  van der Aa,  A.  Pichler,  H.  Ghassemi-Armaki,  K. O.  Findley and  J. G.  Speer: Mater. Sci. Eng. A, 823 (2021), 141569. https://doi.org/10.1016/j.msea.2021.141569
• 30)   A.  Ghatei Kalashami,  C.  DiGiovanni,  M. H.  Razmpoosh,  F.  Goodwin and  N. Y.  Zhou: J. Manuf. Process., 57 (2020), 370. https://doi.org/10.1016/j.jmapro.2020.07.008
• 31)   R.  Ashiri,  Md  A. Haque, C. Ji, M. Shamanian, H. R. Salimijazi and Y. Park: Scr. Mater., 109 (2015), 6. https://doi.org/10.1016/j.scriptamat.2015.07.006
• 32)   D.  Kim,  S.  Hong,  J.  Kang,  Y.  Im and  S.  Kim: Met. Mater. Int., 29 (2023), 135. https://doi.org/10.1007/s12540-022-01214-8
• 33)   D.  Zapico Alvarez,  Y.  Benlatreche,  M.  Brossard,  J.  Giroux,  C.  Kaczynski,  J. M.  Mataigne,  C.  Musik and  P.  Gerkens: BHM Berg-Hüttenmänn. Monatsh., 166 (2021), 546. https://doi.org/10.1007/s00501-021-01166-3
• 34)   A.  Ghatei-Kalashami,  M. S.  Khan,  M.  Lee and  Y. N.  Zhou: Acta Mater., 229 (2022), 117836. https://doi.org/10.1016/j.actamat.2022.117836
• 35)   C.  DiGiovanni and  E.  Biro: Weld. World, 65 (2021), 865. https://doi.org/10.1007/s40194-020-01050-7
• 36)   S. P.  Murugan,  V.  Vijayan,  C.  Ji and  Y.  Park: Weld. J., 99 (2020), 75-s. https://doi.org/10.29391/2020.99.008
• 37)   J.  Zhao,  C.  Ding,  G.  Wu and  J.  Zhang: J. Mater. Res. Technol., 19 (2022), 747. https://doi.org/10.1016/j.jmrt.2022.05.092
• 38)   H.  Peng,  W.  Peng,  R.  Lu,  G.  Wu and  J.  Zhang: J. Alloy. Compd., 806 (2019), 195. https://doi.org/10.1016/j.jallcom.2019.07.232