Enhancement of Mechanical Properties in Dissimilar Resistance Spot Welds between Galvannealed Dual Phase and Al–Si Coated Press Hardening Steels

Abstract

The quality of spot welds between galvanized dual phase steels of 590 MPa (DP590Z) and Al–Si coated press-hardening steels (PHS) of 22MnB5 (PHS1500AS) are determined by welding metallurgy of both base metal and the coating. In this study, extending the dwelling time between two pulses is proposed to suppress splash in the broad process window of welding current and improve mechanical properties of spot welds. The increased welding current enlarges the fusion zone (FZ) size and consequently enhances the strength of welds in both shear and cross tensile tests. Furthermore, martensite with high carbon content and retained austenite near the fusion line was found for the first time in the spot welds in these kinds of steels. The high carbon zone alters the location of broken button and deteriorates the mechanical properties of spot welds. Down-slope pulse is proposed in this study to eliminate carbon enrichment, which improved the mechanical properties of the welds.

1. Introduction

To serve the needs of weight reduction, less CO₂ emission, and meet strict safety standards, the thinner gages of ultra-high strength steels are used in the automobile manufacturing industry.¹⁾ The hot-dip galvanized DP590Z steels are one of the most popular sheet steels for automotive applications owing to the excellent corrosion resistivity by the Zn coating, as well as the lightweight potential, high formability and high strength of 590 MPa.^2,3) The other most widely applied high strength steels are Al–Si coated press-hardening steels of 22MnB5 exceeding 1500 MPa, which are formed into components by hot stamping. The steel blank is heated to the temperature at about 900–950°C for several minutes for full austenitization and then formed and quenched into martensite simultaneously in a die by pressure.⁴⁾ Al–Si coating on PHS can survive the austenitization temperature by forming Fe–Al intermetallic compounds, which prevents oxidization during hot-stamping process and improves the corrosion resistance during service.⁵⁾

It is said that a single automobile has approximate 3000–5000 spot welds in common, which implies that resistance spot welding (RSW) is still a significant joining process. The quality of the spot weld affects the reliability and safety of automobile. Mechanical properties and failure mode of resistance spot welds are the key indexes to evaluate welding spot. Compared with the mild steels and low carbon steels, advance high strength steels (AHSS) usually provide narrow welding current range because of the chemistry, high stiffness, high strength of base metal and complex microstructure.⁶⁾ Furthermore, during welding of coated AHSS, the composition and physical properties (melting point, electrical resistivity, hardness, eta) of the coating are the issues of current range deteriorating. Marashi et al.⁷⁾ demonstrated that FZ size is one of the key factors to determine the maximum force and failure mode in both tensile-shear (TS) and cross-tensile (CT) tests. FZ size is positively correlated to the welding current. Enlargement the welding current range indicates that the increase of maximum welding current before splashing leads to the larger FZ size and better mechanical properties of spot welds.

As far as zinc is concerned, it melts at about 420°C and volatizes at about 910°C, the molten or gasified zinc on the surface and interface trigger explosion molten metal. As a result, pre-heating is regularly applied in RSW of zinc coated AHSS steels for reducing the expulsion tendency by squeezing the molten coating on surface and removing the gasified zinc from interface in order to achieve large FZ size at high welding current.^8,9) In the case of Al–Si coated press hardening steels, the Fe–Al–Si intermetallic compounds with thickness of 30 to 50 μm on the steel surface plays an important role in heat generation and FZ formation in RSW of the PHS steels because of larger electric resistivity compared to the base metal due to the presence of silicon. In addition to the FZ size restricted by the maximum current, the high carbon content of 22MnB5 results in poor toughness of spot welds. The post weld heat-treatment are regularly applied for toughness improvement in the spot welds of PHS1500AS steels. However, the exact origin of the brittleness in the spot weld is not clear yet.

In fact, the joints between DP590Z and PHS1500AS steels in car body are quite common. It is proposed that the interval time between before current pulse and welding current pulse has a significant effect on the expulsion behavior. Meanwhile, the brittleness of the welds is proposed to originate from carbon enrichment near the fusion line. The influence of pre-heating, dwelling time, as well as the change of welding pulse on the microstructure and mechanical properties of the dissimilar resistance spot welds between DP590Z and PHS1500AS are investigated in detail. Finally, optimal welding parameters have been obtained to achieve large FZ size with good mechanical properties.

2. Experiment

2.1. Methods

The geometry and size of specimens for TS and CT tests are described in Fig. 1, according to the JIS3136 standards.¹⁰⁾ The specimens were cleaned with ethanol to remove dirt and oil before welding. A single-phase alternating current resistance spot welder with a 6 mm tip diameter of Cu–Cr dome-radius type electrode was used for spot welding. The cooling water flow rate was 3 L/min. The welds in this study were performed in different welding processes described in Tables 2 and 3. The unit of squeeze, weld, and dwelling times is cycle, where 1 second amount to 50 cycles because of 50 Hz frequency of the alternating current. The weld process window, also known as suitable current range, was between the current at which minimum FZ diameter are equal to 5$\sqrt{t}$ (where t is specimen thickness) and expulsion.¹¹⁾ The TS strength and CT strength were measured using a CMT5105-SANS machine at speed of 2 mm/min at room temperature.

Fig. 1.

Specimen geometries, in mm for tests of (a) TSS and (b) CTS.

Table 2. Welding parameters, the frequency of alternative current is 50 Hz.

	Electrode force /kN	Pulse		Holding Time
		Current	Time
		/kA	/cycle	/cycle
H2	3.5	4.0	30	2
H15		4.0	30	15
H30		4.0	30	30

Table 3. Welding parameters, the frequency of alternative current is 50 Hz.

	Electrode force /kN	Pulse 1		Dwelling Time	Pulse 2		Down-Slope Time	Total welding time
		Current	Time		Current	Time
		/kA	/cycle	/cycle	/kA	/cycle	/cycle	/cycle
D2	3.5	4.0	30	2	6.5–8.0	15	/	47
D15		4.0	30	15	6.5–8.0	15	/	60
D30		4.0	30	30	6.5–9.0	15	/	75
D30-S		4.0	30	30	6.5–9.0	15	20	95

The metallographically polished samples were etched by 4% nital for microstructure analysis, hardness test, FZ size measurements. The FZ size and fracture morphology of welds were observed by optical microscopy (OM). The microstructure of the base metal and the welds were observed by a field-emission scanning electron microscope (SEM) equipped with an electron probe microanalyser (JEOL JXA-8530F) operated at 20 kV accelerating voltage. The elements distribution condition of samples was achieved by the wavelength dispersive spectrometer (WDS). Electron backscatter diffraction (EBSD), operated at 20 kV with a step size of 0.1 μm, was also employed to characterize different microstructural constituents. The micro-hardness test was carried out along a diagonal line with 6 degrees angle and hardness profiles indicate the hardness of base metal (BM), heat-affected zone (HAZ) and FZ, were achieved by the Vickers micro-hardness tester. The micro-hardness tests were carried in load of 500 gf with 10 s dwelling time.

2.2. Materials

The stack materials were made of DP590Z of 1.2 mm thickness and PHS1500AS of 1.2 mm thickness. The chemical composition and tensile properties of base metal sheets are described in Table 1, and the engineering stress-strain curve of different matrix are presented in Fig. 2.

Table 1. Chemical composition and tensile properties of the base metal sheets, where YS, UTS and TEL represent yield strength, ultimate tensile strength and total elongation, respectively.

	Chemical compositions in wt.%								Tensile properties
	C	Si	Mn	Cr	Ti	B	P	S	YS/MPa	UTS/MPa	TEL/%
DP590Z	0.078	0.03	1.76	/	/	/	0.01	0.004	622	409	27
PHS1500AS	0.24	0.15	1.32	0.18	0.026	0.0016	0.004	0.005	1582	1220	7.2

Fig. 2.

Engineering stress-strain curves of DP590Z and PHS1500AS. (Online version in color.)

Figure 3 shows the dual phase of martensite and ferrite constitutes the microstructure of DP590Z, and the microstructure of PHS1500AS samples is fully martensite. The thickness of zinc coating is about 24 μm and Fe–Zn phase is generated between zinc coating and base metal as shown in Fig. 4(a). After hot stamping (Fig. 4(b)), the Al–Si coating reacts with iron to form α-Fe, Fe₂Al₅ and FeAl, which is about 34 μm in thickness. The silicon is distributing in only the α-Fe and FeAl layers since Fe₂Al₅ does not dissolve silicon.

Fig. 3.

SEM micrographs of base metal sheets of (a) DP590Z and (b) PHS1500AS.

Fig. 4.

SEM micrographs and elemental distribution by WDS of (a) zinc layer and (b) Al–Si layer. (Online version in color.)

3. Results and Discussion

3.1. Effect of Dwelling Time between Two Pulses on the Welds’ Properties

In terms of the DP590Z, zinc layer is a key factor in splash during RSW. Therefore, adding a beforehand pulse for melting and removing the coating to restrain splash and broaden the weld process window is necessary.¹²⁾ Many studies have investigated the effect of preheating current and welding time on melting the zinc and its removal.^12,13) Kim et al.¹²⁾ described that the pre-current requires 3 cycles at current of 10 kA to melt zinc and a cooling time of 6 cycles to facilitate removal of the layer. There was no mention of the effect of dwelling time between the pulse. Actually, the dwelling time plays an important role on enlarge welding current range.

Figure 5 shows the microstructure of residual zinc on the interface of the welded sheets achieved by welding parameters with different holding time in Table 2. There is no FZ formed but only a reduction in zinc layer on the interface of welding area due to the less heat input. The thickness of residual zinc on interface of H2 (with 2 cycles hold time) is 11.4±0.6 μm (Fig. 5(a)). As the holding time extend to 15 cycles (H15) and 30 cycles (H30), the thickness of residual zinc on the interface reduce to 7.5±1.0 μm (Fig. 5(b)) and 1.7±0.4 μm (Fig. 5(c)) respectively. Thus, extending the holding time can reduce the residual zinc on the interface of welds effectively.

Fig. 5.

SEM micrographs of residual zinc on the interface of welded achieved by (a) H2, (b) H15 and (c) H30. (Online version in color.)

Furthermore, the current range of different procedures are shown in Table 3. The current range of D2 (with 2 cycles dwelling time) and D15 (with 15 cycles dwelling time) are both from 6.5 kA to 8.0 kA. 1–2 cycles dwelling are commonly applied in practice.^13,14) However, the splash current reached 9.0 kA when the dwelling time was increased to 30 cycles (D30). According to Fig. 6(a), the dwelling time that went up from 2 cycles to 15 cycles evidently influenced neither suitable current range nor size of FZ. In contrast, the suitable current range increased approximately by 1.0 kA because of dwelling time of 30 cycles. During welding, the FZ also increases significantly with the increase of suitable current. The maximum of FZ in D2, D15 and D30 are 6.8 mm, 6.5 mm and 7.5 mm, respectively. FZ size is the most important parameter governing the mechanical properties of spot welds.⁷⁾ Pouranvari and Marashi⁷⁾ reported that there is a proportional correlation between TSS with FZ in expulsion free RSW, the same tendency can be seen in Figs. 6(b), 6(c). The maximum TSS of D2 and D15 was 12.8 kN and 12.6 kN, while the maximum TSS of D30 was 15.3 kN. At the same time, the CTS of welds obtained under the three processes fluctuated between 4 kN and 6 kN. As the FZ experienced shear stress during the TS test, the TS strength reflects the strength of the joint, which increased with the growing of FZ size.⁷⁾ During the CT test, the loading of the notch tip (Fig. 7(b)) may give rise to brittle or semi-brittle fracture at the faying surface, particularly if a brittle microstructure is present in the vicinity of notch tip.⁷⁾ The notch tip is unavoidable at weld of RSW. Thus, the resistance against weld failure in the CT test is determined by toughness of FZ with given size.⁷⁾ As a result, there was no significant improvement in CT test as the FZ size increased from 6.8 mm to 7.6 mm because the CTS of weld with enough fracture toughness has almost reached the maximum value.

Fig. 6.

Characteristics trend with current of welds obtained by D2, D15 and D30: (a) nuggets size, (b) TSS and (c) CTS. (Online version in color.)

Fig. 7.

Simple models describing stress distribution at interface and circumference of a FZ during (a) TS test and (b) CT test.⁷⁾ (Online version in color.)

In the welding process of single pulse, the holding time means the electrode pressure remains on the plates for a certain time after the welding current. Similarly, the time of the electrode pressure remained on the plates between the welding current in multiple pulse was called dwelling time. Therefore, the condition of residual zinc on the interface of welded plates achieved by H2, H15 and H30 could accurately reflect the condition of zinc on the interface of welded plates before the current of pulse 2 in D2, D15 and D30 welding process. Based on the above discussion, sufficient dwelling time is required to achieve a large splash current to enlarge the suitable current range. In view of high hardness of PHS1500AS, molten zinc is extruded by a combination of electrode pressure and sufficient dwelling time. The larger the splash current, the larger is the size of the FZ, and superior are mechanical properties of the welds.

3.2. Effect of Carbon Diffusion by Down-slope Pulse Slope on the Mechanical Properties of Welds

Failure mode of resistance welds is a qualitative measure of mechanical properties.^{15,16,17,18,19)} Basically, spot welds can fail in four distinct modes described as follows: interfacial failure (IF) mode, pullout failure (PF) mode, partial interfacial failure (PIF) mode and partial thickness-partial pullout failure mode (PT-PP) mode.^20,21,22) Generally, the PF mode exhibits the most satisfactory mechanical properties since which means the best energy absorption during failure of welds. However, failure modes are closely related to microstructure distribution and fracture toughness of welds.⁷⁾

The cross-section of the spot weld reveals three distinct structural zones: FZ, HAZ and BM. Owing to the high cooling rate in resistance spot welding, the weld FZ macrostructure shows direction, columnar solidification from the fusion boundary towards the center in all combinations.²³⁾ The microstructure of weld FZ observed in Fig. 8(d) is lath martensite. As the typical hardness curve described in Fig. 9, the hardness of FZ has a slight rise in FZ from DP590Z side to PHF1500AS side. The reason is that the carbon content in martensite of FZ is nonuniform. The initial carbon content of matrix is different in DP590Z (Fe-0.07C-0.03Si-1.76Mn) and PHS1500AS (Fe-0.24C-0.15Si-1.32Mn). The time is not enough for carbon to diffuse sufficiently because of rapid cooling during RSW, which results in different carbon content of martensite, and then leads to a slight change for the hardness of the FZ.

Fig. 8.

SEM micrographs in welding joint of (a) SCHAZ on DP590Z, (b) ICHAZ on DP590Z, (c) UCHAZ on DP590Z, (d) FZ, (e) UCHAZ on PHS1500AS, (f) ICHAZ on PHS1500AS and (g) SCHAZ on PHS1500AS, which obtained by D30 with current of 8.5 kA in pulse 2.

Fig. 9.

Distribution of hardness in welded joint obtained by D30 with current of 8.5 kA in pulse 2. (Online version in color.)

The HAZ can be further divided into three distinct subzones consisting of upper critical HAZ (UCHAZ), intercritical HAZ (ICHAZ) and subcritical HAZ (SCHAZ).⁷⁾ In the UCHAZ, where temperature is high enough for completely austenitizing during welding. In the ICHAZ, the temperature is just between the austenite starting and full austenite points during fast heating of welding, which could be far from the A_c1 and A_c3 temperature measured in very slow heating. In the SCHAZ, the temperature is below austenite starting point in which the martensite in this zone is tempered.⁷⁾ The microstructures from the FZ edge towards the PHS1500AS matrix are as follows: martensite (UCHAZ), martensite and ferrite (ICHAZ) and tempered martensite (SCHAZ). Accordingly, the hardness curve is shown in Fig. 9, which showed that the HAZ hardness increased slightly in UCHAZ, then sharply decreased in ICHAZ and climbed gradually to slightly below the hardness of base metal in SCHAZ. However, the tendency of hardness from the FZ edge towards the DP590Z matrix is different. Though the microstructure of UCHAZ is martensite, the carbon content in martensite is lower than that in FZ, which makes the decreases hardness and then increased slightly to form a platform in UCHAZ. Then the hardness sharply decreased from the UCHAZ to ICHAZ because the microstructure in ICHAZ is martensite and ferrite. The SCHAZ was composed by ferrite and tempered martensite, but the hardness is similar to the hardness of BM as the less tempered martensite. The microstructure described above was observed in Fig. 8.

As shown in Fig. 7(a), the FZ and circumference were suffered shear stress and tensile stress respectively in TS test.⁷⁾ Theoretically, the test sheets should be destroyed on the weaker regions in the softer materials side of DP steel or in the soften zone of ICHAZ of PHS steel if the FZ has large enough diameter and no defect. But the buttons pulled out from the PHS1500AS side, were left over on the DP590Z sheet. Figure 10(a) describes the macroscopic failure circumstance of the sheets after tensile-shear testing. Meanwhile, Fig. 10(b) shows the fracture morphology of the button after testing, lath martensite near the fracture interface indicate that the failure position was close to the fusion line rather than in the ICHAZ.

Fig. 10.

Welds obtained by D30: macroscopic observation after failure in PF mode by (a) photo and (b) OM; weld obtained by D30-8.0 kA: the analysis along the fusion lines of (c) microstructure by OM, (d) phases analysis by EBSD analysis, (e) manganese and carbon distribution by WDS, and (f) hardness profile by micro-Vickers hardness. (Online version in color.)

Figure 10(c) indicates a white area near the fusion line. The white area was predicted to be made up of austenite and martensite. The same sample were reground and polished by electrochemistry for EBSD test to confirm the phase constitutes of the white area. The test location was tagged in figure 10(d), which was grabbed from the red dotted rectangle in Fig. 10(c). As shown in Fig. 10(d), austenite presented in the microstructure, which in the region has a red contrast, while martensite has a blue contrast. The regions, which were not identified by EBSD, are indicated with white contrast. Figure 10(e) shows the variation of carbon content and manganese content from the martensite to retained austenite of the white area. Since the carbon is light element, its quantity cannot be measured accurately, but it is obvious that the carbon content around the austenite increases sharply and is much higher than that in matrix. Morito et al.²⁵⁾ has reported that the dislocation density of martensite increased with increase of carbon content. High dislocation density is known to deteriorate phase identification.²⁴⁾ Therefore, it can be determined that the white area is composed of austenite and high-carbon martensite. Meanwhile, the content of manganese in retained austenite was higher compared with the surrounding phase. The fusion line is the interface of molten metal and unmelted metal after cooling, the carbon and manganese segregation may be easily found near the fusion line as the rapid heating and cooling in RSW. But the mechanism of element segregation forming need to be further study. Figure 10(f) shows the indentation trajectory of microhardness around the high-carbon martensite. The hardness of the martensite with high carbon content is about 750±20 HV. The hardness of martensite with high carbon is about 250 HV more than that of martensite in FZ and base metals. In general, high-carbon martensite is characterized by high hardness and brittleness. Figure 11 shows the microstructure of weld which was stretched to the maximum TS peak load without breaking, cracks were observed in the high-carbon martensite. Therefore, when the specimens endure tensile shear stress, the cracks are inclined and penetrate to the brittle region where stress is concentrated. Eventually, the buttons were pulled out from the fusion line in FZ.

Fig. 11.

Macroscopic observation after the welds were stretched to the maximum TS peak load by D30 with 8.5 kA. (Online version in color.)

Figure 12(a) describes the macroscopic failure circumstance of the welds obtained by D30-S after TS testing. Martensite and ferrite near the fracture interface can be seen clearly in Fig. 12(b), which prove the buttons were broken from the soft zone in the ICHAZ. The distribution of hardness near the fusion line of the weld achieved by D30-S-8.0 kA is relatively uniform in Fig. 12(c). Figure 12(d) shows the variation of carbon content and manganese content near the fusion line. The test location was tagged in Fig. 12(d), which was grabbed from the red dotted rectangle in Fig. 12(c). There was a significant fluctuation in manganese content but not in the carbon content any more. The down-slope current after pulse 2 in D30-S was intended to reduce the cooling speed and to promote the diffusion of carbon, during which manganese as substitutional solutes is not possible to diffuse fast as carbon. It proves that down-slope current can effectively promote carbon diffusion to eliminate martensite with higher hardness and stress concentration. The elimination of high hardness zones can change the failure location of solder joints from FZ to ICHAZ.

Fig. 12.

Welds obtained by process D30-S: (a) macroscopic observation after failure in PF mode by photo; weld obtained by D30-S-8.5 kA after TS test: (b) microstructure by OM; and the analysis along the fusion lines of (c) hardness profile by micro-Vickers hardness; and (d) manganese and carbon distribution by WDS. (Online version in color.)

In CT test, the failure position of welds achieved by D30 and D30-S were both near the fusion line. This result is related to the type of stress applied to the weld. In CT test, tensile stress was applied to the FZ and circumference.⁷⁾ The crack initiates in the notch tip²⁵⁾ and spreads along the thickness direction of the plates whether the martensite with high carbon exists or not. Figure 13 describes the microstructure of weld which was stretched to the maximum CT peak load without breaking, cracks were both observed in the martensite near the fusion line.

Fig. 13.

Macroscopic observation after the welds were stretched to the maximum CT peak load by (a) D30 with 8.5 kA and (b) D30-S with 8.5 kA. (Online version in color.)

As shown in Fig. 14(a), with similar FZ size, the TS strength of welds by D30-S is optimized, compared to that implemented by D30. The extra down-slope pulse in D30-S achieved the TS peak load of 16 kN compared to 14 kN in D30 with same current of 8.5 kA. But there is no significant difference on the CT strength of welds achieved by D30 and D30-S (Fig. 14(b)). Load-displacement curve in TS tests of welds obtained by D30 and D30-S were displayed in Fig. 15(a), the displacements obtained by D30-S were much larger than that by D30. Meanwhile the average fracture absorbing energy of D30-S was about 88 J by calculating the area of the curve compared to about 26 J in D30. In addition to peak loading, displacements and energy absorption were significantly improved as well because of the elimination of stress concentration at the fusion line by depleting carbon enrichment. To be expected, the welds achieved by D30 and D30-S displays similar load-displacement curve (Fig. 15(b)) as the stress concentration eliminating cannot change the failure location in CT test.

Fig. 14.

Characteristics trend with current of welds. obtained by D30 and D30-S in (a) TS tests and (b) CT test. (Online version in color.)

Fig. 15.

Load-displacement curve of welds achieved by D30 and D30-S in (a) TS tests and (b) CT test. (Online version in color.)

4. Conclusions

The major conclusions are as follows:

(1) In resistance spot welding of dissimilar steels with Zn and Al–Si coating, the dual pulse is applied for avoiding expulsion in convention. It has been found the longer dwelling time is beneficial for increasing splash current to promote the broaden current range. The mechanical properties of welds were promoted using a larger welding current when expulsion is suppressed by increasing the dwelling time.

(2) Austenite was found for the first time at the edge of FZ because of the inhomogeneous elemental distribution in the spot welds of conventional automotive steels. The unstable austenite and martensite nearby with relatively high carbon content, exhibits high hardness and brittleness which results in buttons being pulled out from the edge of FZ. Optimizing welding process by adding ‘down-slope’ current contributed to the diffusion of carbon. The mechanical properties of weld achieved by D30-S were significantly enhanced and the buttons of welds were pulled out from the soft zone in the ICHAZ on PHS1500AS or DP590Z side during tensile-shear tests.

Acknowledgements

The research was financially supported by the National Natural Science Foundation (52101128), National Key R&D Program (2018YFE0306102), Postdoctoral Science Foundation (2022M710021), Fundamental Research Funds for the Central Universities (N2007012), 111 Project (B16009), Liaoning Revitalization Talents Program (XLYC1907128) and Postdoctoral Research Fund of Northeastern University (20220202) of China.

References

• 1)   H.  Karbasian and  A. E.  Tekkaya: J. Mater. Process. Technol., 210 (2010), 2103. https://doi.org/10.1016/j.jmatprotec.2010.07.019
• 2)   Y. H.  Jiang,  S.  Kuang,  H. S.  Liu and  H. X.  Teng: Adv. Mater. Res., 1095 (2015), 83. https://doi.org/10.4028/www.scientific.net/AMR.1095.83
• 3)   K. R.  Lee,  S. W.  Lee and  M. B.  Moon: 9th Int. Conf. on Zinc & Zinc Alloy Coated Steel Sheet, 2nd Asia-Pacific Galvanizing Conference (GALVATECH APGALVA 2013), Metallurgical Industry Press, Beijing, (2013), 108.
• 4)   P.  Hein and  J.  Wilsius: Steel Res. Int., 79 (2008), 85. https://doi.org/10.1002/srin.200806321
• 5)   M.  Merklein,  J.  Lechler and  M.  Geiger: CIRP Ann., 55 (2006), 229. https://doi.org/10.1016/S0007-8506(07)60404-1
• 6)   N.  den Uijl and  S.  Smith: Proc. 4th Int. Seminar on Advances in Resistance Welding, (Wels), 4th Int. Seminar on Advances in Resistance Welding, Wels, (2006), 30.
• 7)   M.  Pouranvari and  S. P. H.  Marashi: Sci. Technol. Weld. Join., 18 (2013), 361. https://doi.org/10.1179/1362171813Y.0000000120
• 8)   D. Y.  Choi and  Y. G.  Kim: Text. Res. J., 27 (2009), 7 (in Korean).
• 9)   J.  Zou,  Q.  Zhao and  Z.  Chen: J. Mater. Process. Technol., 209 (2009), 4141. https://doi.org/10.1016/j.jmatprotec.2008.10.005
• 10)  JIS Z 3136: 1999, Specimen dimensions and procedure for shear testing resistance spot and embossed projection welded joints.
• 11)   C-W.  Ji,  I.  Jo,  H.  Lee,  I-D.  Choi,  Y. D.  Kim and  Y-D.  Park: J. Mech. Sci. Technol., 28 (2014), 4761. https://doi.org/10.1007/s12206-014-1043-0
• 12)   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
• 13)   R.  Ashiri,  M.  Shamanian,  H. R.  Salimijazi,  M. A.  Haque,  J-H.  Bae,  C-W.  Ji,  K-G.  Chin and  Y-D.  Park: Scr. Mater., 114 (2016), 41. https://doi.org/10.1016/j.scriptamat.2015.11.027
• 14)   Y.  Chao: Sci. Technol. Weld. Join., 8 (2003), 133.
• 15)   M.  Pouranvari: Can. Metall. Q., 51 (2012), 67. https://doi.org/10.1179/1879139511Y.0000000020
• 16)   M.  Pouranvari,  P.  Marashi,  M.  Goodarzi,  S. M. H.  Sanaee,  A.  Abedi and  S. H.  Abootalebi: Mater. Sci. Technol., 24 (2008), 1506. https://doi.org/10.1179/174328408X262418
• 17)   M.  Pouranvari,  S. P. H.  Marashi and  S. M.  Mousavizadeh: Sci. Technol. Weld. Join., 15 (2010), 625. https://doi.org/10.1179/136217110X12813393169534
• 18)   M.  Pouranvari and  S. P. H.  Marashi: Weld. J., 91 (2012), 303.
• 19)   M.  Khan,  M.  Kuntz and  Y.  Zhou: Sci. Technol. Weld. Join., 13 (2008), 294. https://doi.org/10.1179/174329308X271733
• 20)   M.  Pouranvari: Can. Metall. Q., 50 (2011), 381. https://doi.org/10.1179/1879139511Y.0000000008
• 21)   M.  Pouranvari,  S.  Marashi and  S.  Mousavizadeh: Sci. Technol. Weld. Join., 15 (2010), 625. https://doi.org/10.1179/136217110X12813393169534
• 22)   B. V.  Hernandez,  M.  Kuntz,  M.  Khan and  Y.  Zhou: Sci. Technol. Weld. Join., 13 (2008), 769. https://doi.org/10.1179/136217108X325470
• 23)   E. J.  Seo,  L.  Cho and  B. C.  De Cooman: Acta Mater., 107 (2016), 354. https://doi.org/10.1016/j.actamat.2016.01.059
• 24)   Y.  Kitagawa,  K.  Ikeuchi,  T.  Kuroda,  Y.  Matsushita,  K.  Suenaga,  T.  Hidaka and  H.  Takauchi: J. Mater. Sci., 43 (2008), 12. https://doi.org/10.1007/s10853-007-2150-4
• 25)   C.  Ma,  D. L.  Chen,  S. D.  Bhole,  G.  Boudreau,  A.  Lee and  E.  Biro: Mater. Sci. Eng. A, 485 (2008), 334. https://doi.org/10.1016/j.msea.2007.08.010