Effect of boron content and welding current on the mechanical properties of electrical resistance spot welds in complex-phase steels

Materials and Design 54 (2014) 598–609

Contents lists available at ScienceDirect

Materials and Design journal homepage: www.elsevier.com/locate/matdes

Effect of boron content and welding current on the mechanical properties of electrical resistance spot welds in complex-phase steels Jong Pan Kong a, Tae Kyo Han b, Kwang Geun Chin b, Bong Gyu Park c, Chung Yun Kang a,⇑ a

Dept. of Material Sci. and Eng, Pusan National University, Busan 609-735, Republic of Korea POSCO Ltd., Gumho-dong, Gwangyang-si, Jeonnam 545-090, Republic of Korea c Dept. of Advanced Materials Sci. and Eng., Pukyong National Univ., Busan 608-739, Republic of Korea b

a r t i c l e

i n f o

Article history: Received 31 May 2013 Accepted 28 August 2013 Available online 4 September 2013 Keywords: Complex phase steel Resistance spot welding Boron content Nugget diameter Tensile-shear load

a b s t r a c t When complex phase steel where tensile strength is more than 1 GPa grade is joined by resistance spot welding (RSW) optimum boron (B) content should be chosen to satisfy weldability and mechanical properties. Therefore, in this study, the effect of the B content (0–40 ppm) on the tensile-shear strength of the RSW were investigated. As the resistivity of the base metal was independent on the B content it did not affect to nugget diameter. Regardless of the B content the specimens under 5t1/2 (t = sheet thickness) were fractured at interfacial failure mode. In the low welding current condition (lower than 6.4 kA), measured nugget diameters were smaller than calculated critical nugget diameter regardless of the amount of B addition so that fracture mode was interfacial failure. Pull out failure occurred at the softened zone which was boundary between the base metal and the heat affected zone. Tensile-shear load of the specimen failure at the pull-out mode was increased as the fractured diameter and hardness of the softened zone were increased. Shear load was only dependent on the fractured diameter. The equations to calculate the shear and tensile-shear load were suggested for the specimens fractured at interfacial and pullout failure modes respectively. Correlation coefficients between measured and calculated values of shear and tensile-shear load were 0.98 and 0.97, respectively. Therefore, shear and tensile-shear load of advanced high strength steel joined by RSW could be predicted successfully using the suggested equation. Ó 2013 Elsevier Ltd. All rights reserved.

1. Introduction In order to improve fuel economy of an automobile, the weight of an auto body should be reduced. Lightweight and maintaining strength had trade-off relationship, so there should be much researches to find the optimum combination of them. Recently, due to the strengthened safety regulation for car passengers and pedestrians, installation of the safety device became compulsory. However, this resulted in the increase of weight of the auto body. Therefore, the development of advanced materials with high strength more than 1 GPa and high ductility is essential to overcome this problem. Transformation induced plasticity (TRIP), complex phase (CP), and twinning-induced plasticity (TWIP) steels, etc. which had mixed microstructure with high strength and ductility were produced and make a great contribution to the weigh saving of automotive [1–3]. The CP steel consisted of the mixed microstructure of martensite, retained austenite, bainite and ferrite, and had high

⇑ Corresponding author. Tel.: +82 1083298429; fax: +82 515144457. E-mail address: [email protected] (C.Y. Kang). 0261-3069/$ - see front matter Ó 2013 Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.matdes.2013.08.098

yield strength and good formability so that it was in the spotlight as the material for the auto body. However, the development and application of the CP steel was not sufficient yet [1,2]. On the other hand, in the normal steels boron (B) was known to suppress nucleation of ferrite by segregation in austenite grain boundaries and increase hardenability with the addition of ppm scale [4,5]. This element attracted interest again to secure high strength more than 1 GPa in advanced high strength steel (AHSS). The boron steel for hot-stamping (25 ppm addition of B) which showed the tensile strength of 1.5 GPa after hot stamping was reported in the literature [6]. To produce auto body parts as final product welding process is necessary, and resistance spot welding (RSW) and laser beam welding (LBW) were most widely used. Previous research for this was as follows. Since year 2000, many researches about microstructures and mechanical properties of the weld by RSW and LBW got accomplished in the DP and TRIP steels for automotive which were AHSS with 590–1180 MPa grade [7–14]. Especially, when DP780 and DP980 steels which had high volume fraction of martensite within base metal were joined by RSW softened zone occurred at the boundary of the base metal and HAZ [15–17], and

J.P. Kong et al. / Materials and Design 54 (2014) 598–609

599

Nomenclature TSL ND TSLPF

a FDSZ t

rSZ rBM HvBM HvSZ

tensile shear load (kN) measured nugget diameter (mm) tensile shear load in pull out failure (PF) mode (kN) constant related to the stress distribution diameter of the fractured region in PF mode (mm) thickness of base metal (mm) tensile strength of the softened zone (kN/mm2) tensile strength of the base metal (kN/mm2) hardness of the base metal (Hv) minimum hardness of the softened zone (Hv)

shear strength of the weld nugget (kN/mm2) tensile strength of the weld nugget (kN/mm2) SLPF shear load in interfacial failure (IF) mode (kN) FDN diameter of the fractured region in IF mode (mm) b constant with the notch effect HvN hardness of the weld nugget (Hv) rBM/W.Q tensile strength of the base metal water quenching (kN/ mm2) HvBM/W.Q hardness of base metal water quenching (Hv) NDCri. critical nugget diameter to transit from IF to PF (mm)

sN rN

fracture started at this zone so that the strength of the joint was decreased [18–20]. In addition, many researcher reported the correlation between microstructure/mechanical properties and softening phenomenon on the RSW of these AHSS [14,15,21,22]. Recently, Choi et al. [23] reported that solidification crack and void were formed within nugget in the weld of GA-DP780 and hotstamped Al–Si coated B steels joined by RSW. These defects resulted in the interfacial failure and decreased the strength of the joint. Kim et al. [24] and Hu et al. [25] said that the softening phenomenon occurred at the boundary of base metal/HAZ in the laser weld of CP1180 and CP1000 steels respectively. However, discussions about that phenomenon were not provided sufficiently. Up to now, most of the researches focused on the effect of the process parameters on the microstructure and mechanical properties in the laser and spot welds of DP, TRIP and boron steels. However, in case of the CP steel, the effects of alloying element as well as those effects were hardly found in the literature. On the other hand, Park et al. [26] investigated the effect of the B content on hardness of disk-laser weld in the CP steel and reported that the hardness of base metal and softened zone was increased due to the increase of the martensite volume fraction. This study meant that the B content had an influence on the microstructure and mechanical properties of the resistance spot welds as well as the base metal in the CP steel. However, there was no systematic research about the topic. In this study, in order to obtain optimum B content to satisfy the requirement for weldability and mechanical properties of the resistance spot welds in the CP steel where the tensile strength of base metal was more than 1 GPa, the effects of the B content (0– 40 ppm) and welding current on the tensile-shear load of the resistance spot welds were examined. Correlations between the load, and failure mode and microstructure were also investigated.

the given composition range (0–40 ppm B) the interval of the composition change was 10 ppm from 0B to 10B steels and 15 ppm from 10B to 40B steels. And then, the effect of B content on the mechanical properties was investigated. Spot welding was performed using a PLC-controlled, 120-kVA AC pedestal-type resistance spot welding machine. Welding was conducted using a 45-deg truncated cone RWMA Class 2 electrode with a 6-mm face diameter. The welding currents were varied from 5 to 9 kA, and the welding time, electrode pressure and holding time were fixed to 17 cycles, 4 kN and 40 cycles, respectively. In the present study, the welding parameters were adjusted to avoid expulsion. Tensile-shear tests were performed to evaluate the mechanical performance and failure mode of the spot welds. The samples were prepared due to ANSI/AWS/SAE D8.9M:97 [27]. Fig. 1 shows the test sample dimensions for the tensile-shear tests [28]. The mechanical tests were performed at a cross-head of 5 mm/min with an InstronÒ universal testing machine. The tensile-shear load (measured as the peak point in the load–displacement curve) was extracted from the load–displacement curve. The data points for the tensile-shear load are the average of three measurements. The failure modes of the spot welds specimens were determined by an examination of the fractured samples. Specimens for optical and scanning electron microscopy (SEM) were prepared using standard metallographic practices. Polished specimens were etched with 2% Nital solution and then used to observe the microstructures using optical microscope and SEM. EBSD analysis was carried out using the specimens which were taken from the joint, ground using general SiC paper and polished by 0.04–0.05 lm colloidal silica for 20 min. Vickers mirco-hardness test was performed across the spot welds. A load of 0.2 kgf and a dwell time of 10 s were used during testing.

2. Experimental procedure

3. Results and discussion

The materials used in this study were 1.2 mm thick cold rolled complex steel (CP) sheets containing different amounts of boron (B). Materials used in this study were not fabricated in the lab but supplied by domestic steel manufacturer. Table 1 lists the chemical composition and tensile properties of the base metal. In

3.1. Effect of B content and welding current on tensile strength Generally, the most important factor for the tensile-shear load (TSL) and failure mode is the nugget diameter [11,18,29]. The

Table 1 Chemical compositions and tensile properties of investigated base metals. No.

0B 10B 25B 40B

wt.%

ppm

Tensile properties

C

Si

Mn

Cr

P

S

B

TS (MPa)

YS (MPa)

EL (%)

0.07

0.1

2.1

1.0

200

30

0 10 25 40

840 930 1020 1110

390 470 530 537

21 15 11 9

Fig. 1. Tensile-shear test sample dimension (JIS Z 3136 [28]).

600

J.P. Kong et al. / Materials and Design 54 (2014) 598–609

Fig. 2. Nugget diameter (ND) as a function of B content and welding current.

nugget diameter was defined as the distance between the fusion lines in the cross-section microstructure of the nugget. Fig. 2 shows the change in nugget diameter as functions of the B content and welding current. The marked data (w) are those of the specimens that showed expulsion. The nugget diameter increased with increasing welding current but was less affected by the B content. The addition of B was believed to change the resistivity of the base metal so that the nugget diameter could be increased due to the increased heat input. When the resistivity of the base metals with B content was measured using a 4-point probe [30], the change in resistivity was approximately 0.0011 lO cm/ppm over the composition range of 0–40 ppm. This difference could not have an influence on heat input. Therefore, the nugget diameter was thought to be independent on the B content. Fig. 3 shows the (a) changes of tensile-shear load (TSL) as a function of B contents and welding currents and (b) effect of nugget diameter on the tensile-shear load and failure modes. Fig. 4 shows the representative failure modes of the spot weld during the tensile-shear test. Failure mode of the specimens tested in tensile-shear test was divided into two groups. Fig. 4a presents the typical interfacial failure (IF) mode in fracture surface and Fig. 4b shows the microstructure of the cross-section. Fig. 4c and d presents the fracture surface and microstructure of the cross-section in the specimen that showed typical pull-out failure (PF) mode. In Fig. 3, the open marks showed the values of the IF specimens and the solid marks were PF ones. The specimen welds with a low

current (under 6.4 kA) showed IF mode regardless of the B content and the variation range of the TSL was narrow. On the other hand, the specimen welds with a higher current (over 7.4 kA) showed PF mode, and the TSL increased with increasing welding current and B content. When the welding current was in the range of 7.4–8.4 kA, the increasing gradient of the TSL was smaller than when the welding current was more than 8.4 kA. As shown in Fig. 3b which showed the change of the TSL due to the B content and nugget diameter, the change of the TSL due to the B content could be negligible in the range of the nugget diameter which showed the IF. However, in the range of the nugget diameter which showed PF, the TSL was increased with the nugget diameter. On the other hand, in the same nugget diameter, the TSL was increased with the B content. Generally, the specification for the tensile-shear tests of RSW mild steel and high strength steel for automobiles, ANSI/AWS/ SAE/D8.9-97, defined the minimum nugget diameter where PF was revealed as 4t1/2(where t is the thickness of base metal) [27]. This value is displayed as a broken line () in Fig. 3b. As JIS 3140 specify that it is 5t1/2 in A-class HSS and AHSS [28], vertical dotted line (——) where the minimum nugget diameter was 5t1/2 was drawn to compare with values obtained in this study. Also, horizontal solid line (–) was drawn where the minimum tensileshear load specified in the JIS 3140 A-class was 8.78 kN [28]. From the result in Fig. 3b, regardless of B content the IF occurred under the 5t1/2 (5.48 mm) and PF was over 6.2 mm of the nugget diameter. This means that the IF/PF transition critical nugget diameter should be 5t1/2 rather than 4t1/2 in the specimen used in this study. In addition, the TSL of all welds satisfied minimum TSL values (8.78 kN) specified in the JIS 3140 A-class [28], even though the specimen was fractured in IF mode. To ensure the collision safety of motor vehicles, the impact absorbed energy is very important and was reported to have a correlation with the failure mode [18,19]. Fig. 5 shows the relationship between the nugget diameter and tensile absorbed energy with the B content. The amount of absorbed energy was calculated by measuring the area under the load–displacement curve up to peak loads. From the results in Fig. 5, the specimens that exhibited IF mode showed little change in the tensile absorbed energy due to the increase in nugget diameter and B content; the value was low (5–10 kN mm). On the other hand, in the specimens showing PF mode, the absorbed energy increased with increasing nugget diameter, and was 3–5 times higher than those with IF mode. Moreover, in the nugget diameter range that resulted in PF mode, the tensile absorbed energy decreased with increasing B content at the same nugget diameter. In order to secure the collision safety, the joint of AHSS welded

Fig. 3. (a) Tensile-shear load as a function of B contents and welding currents and (b) effect of nugget diameter on the tensile-shear load and failure modes.

J.P. Kong et al. / Materials and Design 54 (2014) 598–609

601

Fig. 4. Typical failure mode observed in this study. (a and b) interfacial failure and (c and d) pull-out failure.

Generally, the position where the specimen welds failed in PF mode affects the TSL. Therefore, this needs to be clarified. However, in Fig. 4d which showed the cross-section microstructure of the specimen welds failed in PF mode, the location was not clearly

defined. Accordingly, the area marked X, which was the lower part of Fig. 4d, was observed carefully. Fig. 6 is enlarged optical micrograph of the X-marked area of Fig. 4d. As shown in Fig. 6, failure (arrow) was initiated in the boundary between the base metal and HAZ. Fig. 7 shows typical hardness profile near the base metal and fusion line (FL) of the 10B steel weld. The hardness profile in Fig. 7 showed that softened zone was existed near the boundary between the base metal and HAZ (sub-critical HAZ; heated area under Ac1 temperature). Some researches reported that the softened zone occurred on the base metal/HAZ boundary in the specimen welded by RSW where the tensile strength was over 780 MPa grade AHSSs (DP and CP steels) and the failure was initiated in this zone [18–20]. From the observation of the microstructure near the failure (Fig. 6), hardness profile (Fig. 7) and former researches, the PF was thought to be initiated in the region where the hardness was the lowest in the softened zone. The right side of the lower plate also fractured at the same mode. All specimens that showed PF mode exhibited a similar behavior, i.e. failure was started at the location where the hardness was the lowest in the softened area. Therefore, the hardness of the softened area is an important parameter affecting the TSL of the welds. The hardness of the base metal is believed to be related to the B content, and the softening phenomenon is associated with the welding current. Therefore, the effect of the B content on the base metal hardness and microstructure was examined first. Fig. 8 shows that change in hardness of the base metal increased with increasing B content. The microstructures of the

Fig. 6. Enlarged optical micrographs of the area denoted by X in Fig. 4d. (10 B steel welds, welding current: 7.4 kA).

Fig. 7. Typical hardness profiles of 10B steel welds. (welding current : 7.4 kA).

Fig. 5. Absorption energy as a function of B contents and welding current.

by RSW must be fractured in the PF. Therefore, the regulation for the minimum nugget diameter should be revised. 3.2. Softening phenomenon in the fracture surface with PF mode

602

J.P. Kong et al. / Materials and Design 54 (2014) 598–609

Fig. 8. Hardness of base metal as a function of B contents.

specimens where 0 and 40 ppm B was added were compared to examine the correlation between the B content and the hardness and microstructure. Fig. 9a and b shows the optical microstructure of 0B steel (a) and 40B steel (b), which were color-etched using Lepers’s solution. Fig. 9c and d presents SEM images of the X- and Y-marked areas in Fig. 9a and b, respectively. Ferrite (F) was distinguished as blueviolet, Bainite (B) was brown, and Martensite/Retained Austenite (MA) was white. The microstructure of the base metal was composed of ferrite, bainite and MA regardless of the B content. In addition, the size of martensite and its fraction increased with increasing B content. To examine the effect of the B content on the volume fraction of each phase, 10 SEM images were taken in each specimen and the area fraction was measured using Image Pro Plus software. Fig. 10 shows the change in the volume fraction of each phase with the B content. In this result, the volume fractions of martensite and bainite increased linearly with increasing B content, whereas the ferrite faction decreased linearly. The increase in the hardness of the base metal increased with increasing B content

Fig. 10. Change of the volume fraction of each phase as a function of B content. F, ferrite; M-A, martensite/retained austenite; B, bainite.

was attributed to the decreasing volume fraction of the soft phase (ferrite), and increasing hard phase (martensite and bainite). In order to clarify the reason why the microstructure was significantly changed by the addition of trace amount of B(40 ppm), CCT curve due to B content was drawn using JMatPro software (version 5). The resulting curves (sold line) where the parameters of heat treatment cycles used to fabricate the steels were assigned to the software are shown in Fig. 11. In the curves, open symbols were for 0B steel and closed ones were for 40B steel. As shown in the curves, Ms temperature was not affected by the B addition (40 ppm) while initiation time for ferrite and bainite was shifted to the longer side. In the heat treatment cycle used to fabricated the steels, cooling curve for the 0B steel passed through ferrite nose white that for 40B steel did not. On the other hand, Melloy et al. [4] and Seol et al. [5] reported that the formation of segregated ferrite along the austenite grain boundary was suppressed by the addition of B. In addition,

Fig. 9. (a and b) Optical micrographs of base metal with different B content and (c and d) SEM micrographs of the X-and Y-marked areas in Fig. 9a and b, respectively. Etchant, LePera’s tint; F, ferrite (blue–violet), B, Bainite (Brown); and M–A, martensite/retained austenite (white). (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

603

J.P. Kong et al. / Materials and Design 54 (2014) 598–609

Generally speaking, the degree of softening in the sub-critical HAZ was dependent on the retention time near the Ac1 temperature (eutectic temperature) during welding where the martensite could be decomposed [11]. And, the retention time might be extended since the heat input was increased with welding current. The higher the welding current was the longer the retention time at the temperature (Ac1 temperature) where tempering was in progress was. As martensite was decomposed during this period, the hardness of the softened zone seemed to be decreased as shown in Fig. 12. This result was well consistent with that of Pouranvari et al. [11]. 3.3. Effect of the B content and welding current on the tensile-shear load in PF mode Fig. 11. CCT curve of 0B steel and 40B steel calculated using JmatPro v-5 software. Dotted line: 0 B steel, solid line: 40B steel, Bs: bainte (0.1%), Fs: ferrite (0.1%), Ps: pearlite, and Ms: martensite start temp.

Nobuyuki et al. [31] also reported that the cooling curve for ultra high strength steel did not ferrite nose in spite of the increase of B addition so that the strength was increased due to the increase of bainite volume fraction. From the above results (Figs. 9–11) and previous researches [4,5,31] and martensite volume fractions were increased and that of ferrite was decreased as the B content increased even though the steel followed the same heat treatment cycle. It resulted in the increase of the hardness of the base metal. Fig. 12 shows the change in the minimum hardness in the softened zone with the welding current and the hardness of the base metal with the B content. The minimum hardness of the softened zone was reduced much more drastically than that of the base metal, and the rate of hardness reduction with increasing welding current was similar, regardless the B content . This study examined why the softening phenomenon occurred in the sub-critical HAZ near the base metal during welding and the effect of the welding current on the softening phenomenon. Fig. 13 shows typical SEM images of the base metal (a) and softened zone (b) in welded 10B steel. In the softened zone (Fig. 13b), the microstructure was composed of a mixture of tempered martensite (T.M) and untransformed martensite (M), where it existed in the base metal. Based on this observation, the softening phenomenon was attributed to a transformation of martensite in the sub-critical HAZ near the base metal to tempered martensite due to the heating of Ac1 (eutectoid temperature) by heat input. This softening phenomenon was reported to appear in the spot welds in DP780 and DP980 steels [15–22].

Fig. 12. Hardness of base metal and minimum hardness of softened zone (HvSZ) as a function of B content and welding current.

When resistance spot weld was fractured in the PF mode outside the nugget, several empirical equations to predict the tensile-shear load (TSL) were suggested. That is, Eq. (1) was suggested by Heuschkel et al. [32], Eqs. (2) and (3) was by Kabasawa et al. [33] and Eq. (4) was by Pouranvari et al. [11]. Eqs. (1) and (2) explained that the TSL was proportional to the plate thickness (t), strength of the base metal (rBM) and nugget diameter (ND). Eq. (3) is an empirical formula of the model where the stress distribution acting on the circular section was considered, and Eq. (4) denotes the circular section without considering the stress distribution. Eqs. (3) and (4) used the fractured diameter (FD) as the parameter affecting the TSL, particularly the strength of the fractured part (rFL) used in Eq. (4):

TSLPF ¼ A  ND  t  rBM 0:72

TSLPF ¼ 17:52  ND

ð1Þ tr

0:81 BM

ð2Þ

TSLPF ¼ 1:95  t  rBM ð1 þ 0:0059  ELBM Þ FD ¼ 2:05  t  rBM ð1 þ 0:0059  ELBM Þ  ðND þ 2:09Þ TSLPF ¼ p  ND  t  rFL or TSLPF ¼ p  FD  t  rFL

ð3Þ ð4Þ

where TSLPF is tensile-shear load in PF mode, A is a materials dependent coefficient (A = 2.5–3.0), t is thickness, rBM is tensile strength of base metal, ND is nugget diameter, FD is fractured diameter (Button diameter), ELBM is elongation of base metal, rFL is tensile strength of the pull out fracture location. In this study, when the constant thickness of the steel plate and fracture mode of the welds were considered, the parameters affecting to the TLS were the strength of the fractured part (rFL) and fractured diameter (FD). In particular, as the strength of the fractured part is unknown, the correlation with TSL was examined using the hardness of the fractured part, i.e. minimum hardness of the softened zone. Fig. 14 shows the change in tensile-shear load (TSL) due to the B content and welding current, and variation of the TSL due to the fractured diameter (a) and minimum hardness of the fractured part (softened area hardness) (b). The fractured diameter was defined as the distance between the locations which showed the lowest hardness values (twice the distance from the nugget center to the location which showed the lowest hardness values) in the weld section microstructure because PF was initiated at the location in the softened zone where the hardness was the lowest which was observed in the failure section microstructure (Fig. 6). It was measured using Image Pro Plus software. As shown in Fig. 14a, the fractured diameter was increased with the welding current regardless of the B content in all specimens, and then, the TSL was also increased. At the same welding current, the TSL was increased with the B content. In case of the change of TSL due to the minimum hardness of the softened zone (Fig. 14b), the hardness of the softened zone was decreased as the welding current increased regardless of the B addition in all specimens

604

J.P. Kong et al. / Materials and Design 54 (2014) 598–609

Fig. 13. Typical SEM micrographs of (a) base metal, (b) softened zone in 10B steel welds. Etchant: nital 4%; M: martensite, and T.M: tempered martensite [welding current: 7.4 kA].

Fig. 14. (a) Tensile-shear load with fractured diameter and (b) minimum hardness of softened zone as a function of B content and welding current in pull-out fractured specimen.

but the TSL was high. This could be rationalized by the following explanation. In the range of the welding current where the pull out failure occurred, the increasing rate of the fractured diameter (about + 8.8%/kA) was higher than the decreasing rate of the softened zone hardness (about 2.5%/kA) as the welding current was increased so that tensile-shear strength was high. Therefore, the fractured diameter was more dominant parameter than the hardness of the softened zone on the TSL in the given range of the welding current. On the other hand, Fig. 14a showed that TSL was increased as the B content was increased even though the fractured diameter was hardly changed at the same welding current. This seemed to be because the hardness of the softened zone was increased as the B addition was increased. Based on the results inferred from Fig. 14, the TSL of the spot welds in CP steels with different B content was examined when the specimen was fractured in PF mode during the tensile-shear test. As mentioned above, the TSLPF was expressed as Eq. (5) when the diameter of the fractured part was proportional to the strength of the fractured part and the constant related to the stress distribution is given as a:

TSLPF ¼ a  FDSZ  t  rSZ

ð5Þ

where TSLPF is tensile-shear load in PF mode, a is a constant, FDSZ is the fractured diameter (actually, twice the distance from the center of the nugget to the part that showed the minimum hardness) in PF mode, t is the base metal thickness, and rSZ is the tensile strength of the softened zone. Generally, the tensile strength is proportional to the hardness, and the tensile strength of the softened zone can be predicted with

the ratio between the hardness of the base metal and softened zone. Therefore, TSLPF could be expressed by the following equation:

TSLPF ¼ a  FDSZ  t  rBM  ðHvSZ =HvBM Þ

ð6Þ

where rBM is the tensile strength of the base metal (kN/mm2), HvBM is the hardness of the base metal (Hv), and HvSZ is the minimum hardness of the softened area (Hv). On the other hand, for all specimens, the correlation between the nugget diameter (ND) and fractured diameter (FDSZ) was analyzed by regression analysis and is expressed as:

FDSZ ¼ ND þ 1:5

ð7Þ

If Eq. (7) is assigned to Eq. (6), the equation to calculate the TSLPF can be expressed as:

TSLPF ¼ a  t  rBM  ðHvSZ =HvBM Þ  ðND þ 1:5Þ

ð8Þ

The constant a related to the stress distribution was calculated by transformation of Eqs. (8) and (9). That is, each property (the thickness of the base metal, tensile strength, hardness, nugget diameter and TSL) of the weld and the base metal due to the B content and welding current were assigned to Eq. (9), and then, a was calculated by least square method. In the given range of the B content and the welding current, its average was about 2.29.

a ¼ TSL=ðt  rBM  ðHvSZ =HvBM Þ  ðND þ 1:5ÞÞ

ð9Þ

where a was constant related to the stress distribution, rBM tensile strength of the base metal (kN/mm2), Hv.BM hardness of the base

J.P. Kong et al. / Materials and Design 54 (2014) 598–609

Fig. 15. Correlation between measured tensile-shear load (TSLPF) and calculated TSLPF.

metal (Hv.), Hv.SZ minimum hardness of the softened zone (Hv.), ND nugget diameter (mm), and TSL measured TSL (=peak load) (kN). Fig. 15 shows correlations between measured and calculated tensile-shear loads. The calculated values were obtained from Eq. (8) where properties of the weld and base metal due to the B content and welding current were assigned. The measured values were expressed as closed mark. Open mark data was taken from the previous researches of Pouranvari et al. (DP780, DP980, 1.5mmt) [11], Sun et al. (DP800, 1.6mmt) [18] and Khan et al. (DP780, 1.15mmt) [34]. These data including that measured in this study were used to analyze the correlation between the calculated and measured TSL. The correlation coefficient was 0.97. Therefore, in the tensile-shear test of the DP steel and CP steels, which contained 0–40 ppm B welded by RSW, the TSLPF of the part failed with PF mode could be predicted by Eq. (8).

3.4. Effect of the B content and welding current on the shear load in IF mode The shear load (SLIF) of an IF mode was expected to be affected by the shear strength and nugget diameter because IF occurs along the joint interfaces of nuggets in a tensile-shear test. Although the shear strength of a nugget is not known, the shear strength is expected to be proportional to the hardness of the nugget. Therefore, the relationship between the hardness and nugget diameter, which may have an effect on the SLIF, was reviewed. Fig. 16 shows the correlations among the SLIF, nugget diameter, and hardness of a welded nugget in which an IF occurred. Although the nugget diameter increased with increasing welding current, the B content has almost no effect. Moreover, the SLIF increases with increasing nugget diameter. On the other hand, while the hardness of the nugget was almost constant, irrespective of the B content and welding current, the SLIF varies substantially with the welding current. This shows that the SLIF is dependent only on the nugget diameter in the case of IF mode. As mentioned above, there was almost no difference in the specific resistance in the B composition range of 0–40 ppm, and the heat input was dependent only on the current such that the nugget diameter increases with increasing current increases. Hence, SLIF increases. The reasons for the relatively constant hardness of the nugget with B content and welding current are reviewed. Fig. 17 shows SEM images of the nugget at different B content ((a and b) 0B steel, (c and d) 40B steel) and welding current ((a

605

Fig. 16. Relationship between shear load and nugget diameter (ND), and that between average hardness of nugget (Hv) with B content and welding current in interfacially fractured specimen.

and c) 5.4 kA, (b and d) 6.4 kA). The microstructure of the nugget was determined to be full martensite regardless of the B content and welding current. The reason why the nugget microstructure was full martensite regardless of the B content and the welding current could be explained using the CCT curves in Fig. 11. Generally, in the resistance spot weld of the steel plate whose thickness was 1.2 mm, cooling rate in the nugget was reported to be 6000 °C/s when calculated using FEM [35]. The base metal was heated rapidly up to 850 °C using high frequency induction furnace and water cooled, and then, the cooling rate was measured using thermocouple. It was about 600 °C/s. In Fig. 11 cooling rate of the resistance spot weld was drawn as dashed line (— —) and that of the water cooled specimen was dotted line (——). As shown in Fig. 11, only a martensitic transformation occurred without passing through a bainite transformation curve (Bs) regardless of the B content because the cooling rates of RSW and water cooling are quite high. Accordingly, it was estimated that the hardness of the nugget is constant because all nuggets have a full martensite microstructures because of the rapid cooling characteristics of the RSW within a given composition range of B. 3.5. Prediction of the shear load in IF mode Generally, SLIF in IF mode can be expressed as a product of the shear strength of the nugget and the area of the failed cross section. In the uniaxial tension test of a thin plate, the tensile strength (r) and shear strength (s) had the following relation, i.e. s = 0.5r due to the Tresca criterion [36]. Therefore, the SLIF can be expressed as:

SLIF ¼ sN  p  ðFDN =2Þ2 ¼ 0:5  rN  p  ðFDN =2Þ2

ð10Þ

where SLIF is the shear load in IF mode, sN is the shear strength of the welds nugget, FDN is fractured diameter in IF mode, and rN is tensile strength of nugget. The nugget near the interface was observed carefully to estimate the area of the failure location. Fig. 18a shows the bonding state of a nugget of 10B steel welds with a current of 6.4 kA, and (b) is the inverse pole figure color map obtained by EBSD measurements for the X-marked area of Fig. 18a. As shown in Fig. 18b, the base metal outside the nugget in both sides could be divided into the bonded (corona bond) and unbonded regions. Because IF proceeded along the interface of the joint, the fractured diameter

606

J.P. Kong et al. / Materials and Design 54 (2014) 598–609

Fig. 17. SEM micrographs of welds nugget as a function of B content and welding current. (a) 0B steel-5.4 kA, (b) 0B steel-6.4 kA,(c) 40B steel-5.4 kA, and (d) 40B steel-6.4 kA.

Fig. 18. (a) Optical micrograph near corona bond in 10 B steel spot welds and (b) inverse pole figure color map by EBSD for X-marked area of Fig. 18a. [welding current: 6.4 kA]. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

(FDN) could be expressed by the sum of the nugget diameter (ND) and corona bond length (CBL) of both sides, i.e. FDN = ND + CBL. In addition, the shape of the cross section in Fig. 18 indicates that when the specimen fractured by SLIF, the formation of cracks is accelerated at the triple point, P, of the HAZ, and both sides of the base metal act as a sharp notch so that the shear stress is thought to have decreased. Considering the notch effect, Eq. (10) could be rewritten as:

SLIF ¼ b  0:5  rN  p  ðFDN =2Þ2

ð11Þ

where b is the proportional constant of the notch effect. FDN can be measured from the nugget cross section, but the real tensile strength (rN) of the nugget cannot be measured. This was obtained by conversion using the following method. As shown in Fig. 18, the microstructure of the nugget was full martensite regardless of the B content and welding current. When each base metal was heated to temperatures over AC3 (801 °C) and water quenched, it was assumed that the tensile strength of the specimen would have a correlation with the tensile strength of the nugget. Fig. 19 shows SEM images of the base metal water quenched directly from 850 °C after rapid heating to that temperature. All samples were full martensite. From the results in Figs. 17 and 19, the microstructures of the nugget and water cooled specimen were known to be full martensite. However, prior austenite grain size (G.SP.A) would be not the

same because transformation mechanism was different. The nugget microstructure was formed by phase transformation of the base metal which was melt by resistant heat and solidified rapidly so that G.SP.A could be relatively big. But, water cooled specimen was austenitized at 850 °C which was near AC3 temperature (801 °C) so that G.SP.A was thought to be small. Therefore, G.SP.A microstructures of the nugget and water cooled specimen were revealed by etching with NaOH solution (NaOH 27 g + picral 2.7 g + H2O 125 mL) in the 10B steel welded with the welding current of 6.4 kA, and their size was measured by line method (ASTM E112-96) [37] and compared each other. The G.SP.A of the water cooled specimen was about 12.1 lm and lower than that of the nugget (about 84.6 lm). 40B steel showed similar trend. This difference of the G.SP.A might result in the difference in the tensile strength of the nugget and water cooled base metal. Therefore, the tensile strength of the nugget (rN) was obtained by multiplication of hardness ratio of the nugget/water cooled base metal to the tensile strength of the water cooled base metal. The shear load of the specimen failed in the interfacial failure mode could be expressed as:

SLIF ¼ b  0:5  rBM=W:Q  ðHvN =HvBM=W:Q Þ  p  ðFDN =2Þ2

ð12Þ

where SLIF is the shear load in IF mode (kN), b is a constant, rBM/W.Q is the tensile strength of the water-quenched base metal (kN/mm2), HvN is the hardness of the nugget (Hv), HvBM/W.Q is the hardness of

607

J.P. Kong et al. / Materials and Design 54 (2014) 598–609

Fig. 19. Typical SEM micrographs of water quenched base metal as a function of B content. (a) 0B steel and (b) 40 steel.

welded CP steel with the given range of the B content (0–40 ppm) was subjected to the tensile-shear test.

Table 2 Comparison of experimental shear load and predicted shear load by Eq. (12). No.

Current (kN)

Experimental SL (kN)

Predicted SL (kN)

0B

5.4 6.4

8.74 13.55

8.37 13.80

3.6. IF to PF transition nugget diameter

10B

5.4 6.4

8.81 13.65

8.64 13.90

25B

5.4 6.4

8.91 13.88

8.56 14.10

40B

5.4 6.4

8.99 13.96

8.68 14.05

The standard for the spot welds of the steel sheet of automotive (ANSI/AWS/SAE/D8.9-97) defined the minimum nugget diameter as 4t1/2 or 5t1/2(t = the thickness of the base metal) for PF to occur [27,28]. The minimum nugget diameter has the same meaning as the critical nugget diameter (NDCri.), where the transition from IF mode to PF mode. NDCri. can be calculated from the correlation between Eq. (8) for TSLPF in PF mode and Eq. (12) for the SLIF in IF mode. Therefore, the equation for TSLPF is the first degree function of the nugget diameter (ND) and SLIF was second degree function of the ND so that NDCri. is the point of intersection of two plots, i.e. ND vs. SLIF and ND vs. TSLPF. In order to plot the changes of the SLIF and TSLPF due to the ND, mechanical properties of the base metal and weld with B contents and welding current are listed in Table 3. In particular, an increase in the amount of B added resulted in an increase in the martensite volume fraction in the base metal, the tensile strength (óBM) and hardness of the base metal (HvBM), and hardness of the softened zone (HvSZ). On the other hand, the tensile strength (rBM/W.Q) and hardness of the water-quenched base metal (HvBM/W.Q), and the hardness of the nugget (HvN) were relatively unchanged because the microstructures were full martensite regardless of the B contents. Therefore, the average was assigned to Eq. (12). Fig. 20 shows a plot of the SLIF and TSLPF vs. ND that was produced at welding currents of 6.4 kA (a) and 7.4 kA (b) after the data in Table 3 was assigned to Eqs. (8) and (12). The point of intersection, (s), of the SLIF curve and each TSLPF line was NDCri., and the point, (j), was the ND measured in this study. Table 4 lists the values of the NDCri. due to the B content and welding current, which were calculated and measured. For the specimen welded at a welding current of 6.4 kA, b was assigned to 0.89, whereas for the specimen welds at 7.4 kA, b = 1 because the notch was on the side of the base metal not on the interface, as shown in Fig. 6.

the water-quenched base metal (Hv) and FDN is the fractured diameter in IF mode. Proportional constant (b) with notch effect was obtained from the transformed Eqs. (12) and (13). That is, properties of the base metal and weld (tensile strength, hardness, fractured diameter and shear load) due to the B content and welding current were assigned to Eq. (13) and least square method was used to analyze them. In the given ranges of the B content and welding current the average b was about 0.89. The fact that the b was lower than 1 meant that the notch stress had an influence on it. Besides, the b seemed to be changed slightly due to the steel composition and mechanical properties.

b ¼ SL=ð0:5  rBM=W:Q  ðHvN =HvBM=W:Q Þ  p  ðFDN =2Þ2 Þ

ð13Þ

where SL was measured shear load (kN) in the specimen fractured at IF mode. Table 2 lists the calculated and measured shear load. The calculated values were obtained by calculation where the properties of the base metal and weld due to the B content and welding current were assigned to Eq. (12). And measured values were taken in this study. The calculated and measured values were well consistent (R2 = 0.98). Therefore, Eq. (12) could be used to predict the shear load of the specimen fractured at IF mode when the resistance spot

Table 3 Mechanical properties of steels as a function of B content and welding current. No.

0B 10B 25B 40B

rBM (MPa) 850.2 930.3 1020.4 1110.6

HvBM (Hv)

247.8 273.7 321.0 347.0

rBM/W.Q (MPa) 1325.6

HvBM/W.Q (Hv)

384.0

6.4 kA

7.4 kA

HvSZ (Hv)

HvN (Hv)

HvSZ (Hv)

HvN (Hv)

236.1 253.1 281.6 298.2

378.1

231.2 245.1 275.3 290.5

378.0

where rBM , tensile strength of base metal; HvBM , hardness of base metal; rBM=W:Q , tensile strength of water quenched base metal; HvBM=W:Q , hardness of water quenched base metal; HvSZ , hardness of softening zone; and HvND , hardness of nugget.

608

J.P. Kong et al. / Materials and Design 54 (2014) 598–609

Fig. 20. Tensile-shear load vs. critical nugget diameter and measured nugget diameter as a function of B content and welding current in IF and PF modes. (a) 6.4 kA and (b) 7.4 kA.

Table 4 Critical nugget diameter (NDCri.) to ensure the PF mode and measured nugget diameter (ND) as a function of B content and welding current. No.

0 10 25 40 Mode

6.4 kA

7.4 kA

NDCri. (mm)

ND (mm)

NDCri. (mm)

ND (mm)

6.16 6.40 6.60 7.22 IF

5.38 5.38 5.42 5.40

5.40 5.80 5.90 6.20 PF

6.21 6.20 6.21 6.22

From Fig. 20a, TSLPF (sold line) increased linearly with increasing ND, and the absolute value of TSLPF increased with increasing B contents. On the other hand, SLIF (dotted line) increased in a parabolic manner with increasing nugget diameter regardless of the B content. NDCri. increased with increasing amount of B. The specimen welds at 7.4 kA showed a similar trend (Fig. 20b). The specimen welds at 6.4 kA fractured with IF mode because the ND was smaller than NDCri. (ND was located at the left side of the NDCr on the SLIF line.) regardless of the amount of B added. On the other hand, for the specimen welds at 7.4 kA, the joint was fractured in PF mode because all the values of the ND were larger than those of the NDCri. (ND was located at the right side of the NDCri. in the TSLIF line.). 4. Conclusions The effects of the boron (B) content and welding current on the mechanical properties of the resistance spot welded in CP steel were examined and the following conclusions were made: (1) The nugget diameter (ND) and fractured diameter (FD) was proportional to the welding current. B content (0–40 ppm) had little influence on the ND and FD because it did not change the resistivity. (2) In the tensile-shear test failure mode was divided by two, that is, interfacial failure (IF) and pull out failure (PF). The PF was occurred in the softened zone (the place where the hardness was the lowest in the weld) at the base metal/ HAZ boundary. As the B content was increased the volume fraction of martensite within the base metal increased so that the hardness of the base metal and softened zone was increased. The fractured diameter was increased with the welding current, but the B content had little influence on it.

(3) The tensile-shear load (TSLPF) of the specimen fractured at PF mode was increased with the fractured diameter and the hardness of the softened zone. The correlation between the calculated TSLPF values taken from the equation to predict TSLPF taken in this research and measured one was analyzed and the correlation coefficient (R2) was 0.97. This meant both of them are well correlated. (4) The nugget microstructure was full martensite and little changes due to the B content and welding current. The shear load (SLIF) was dependent only on the fractured diameter. The correlation between the calculated SL values taken from the equation to predict SLIF taken in this research and measured one was analyzed and the correlation coefficient (R2) was 0.98. This meant both of them are well correlated. (5) The IF occurred during welding with low current (under 6.4 kA) where the nugget diameter (ND) was smaller than 5t1/2 (5.48 mm) regardless of the B content. In the CP steel with given range of the B content the critical nugget diameter (NDCri.) for transition from the IF to the PF was known to be 5t1/2 rather than 4t1/2 (t = the thickness of the base metal). In the low current welding (under 6.4 kA) the predicted ND was smaller than the NDCri. so that the IF occurred regardless the B addition.

Acknowledgment This work was supported by the National Research Foundation of Korea (NRF) grant funded by the Korea Government (MSIP) (No. 2012R1A5A1048294). References [1] Kuziak R, Kawalla R, Waengler S. Advanced high strength steels for automotive industry. Arch Civ Mech Eng 2008;8(2):103–17. [2] ULSAB-AVC consortium. ULSAB-AVC technical transfer dispatch# 6 – body structure materials; 2001. [3] ULSAB-AVC consortium. New high strength steels applied to the body structure of ULSAB-AVC; 2001. [4] Melloy GF, Summon PR, Podgursky PP. Optimizing the boron effect. Metall Trans 1973;4(10):2281–9. [5] Seol J B, Lim NS, Lee BH, Renaud L, Park CG. Atom probe tomography and nano secondary ion mass spectroscopy investigation of the segregation of boron at austenite grain boundaries in 0.5 wt.% carbon steels. Met Mater Int 2011;17(3):414. [6] Karbasian H, Tekkaya AE. A review on hot stamping. J Mater Process Technol 2010;210(15):2104–5. [7] Yang HG, Zhang YS, Lai XM, Chen Guanlong. An experimental investigation on critical specimen sizes of high strength steels DP600 in resistance spot welding. Mater Des 2008;29(9):1679–84.

J.P. Kong et al. / Materials and Design 54 (2014) 598–609 [8] Son JW, Nam DG, Kim DC, Park YD. Effect of Paint Baking on the Strength and Failure of Spot Welds for 780 TRIP Steels. J Korean Weld Soc 2010;28(2):186–93. [9] Sharma RS, Molian P. Yb:YAG laser welding of TRIP780 steel with dual phase and mild steels for use in tailor welded blanks. Mater Des 2009;30(10):4146–55. [10] Pouranvari M, Marashi SPH. Failure mode transition in AHSS resistance spot welds. Part I. Controlling factors. Mater Sci Eng A 2011;528(29):8337–43. [11] Pouranvari M, Marashi SPH, Safanama DS. Failure mode transition in AHSS resistance spot welds. Part II. Mater Sci Eng A 2011;528(29):8344–52. [12] Yu JY, Shim JH, Rhee SH. Characteristics of resistance spot welding for 1 GPa grade twin induced plasticity steel. Mater Trans 2012;53(11):2011–8. [13] Xu W, Westerbaan D, Nayak SS, Chen DL, Goodwin F, Zhou Y. Tensile and fatigue properties of fiber laser welded high strength low alloy and DP980 dual-phase steel joints. Mater Des 2013;43:373–83. [14] Marya M, Wang K, Hector LG, Gayden JX. Tensile-shear forces and fracture modes in single and multiple weld specimens in dual-phase steels. J Manuf Sci Eng 2006;128(1):287–98. [15] Baltazar Hernandez VH, Nayak SS, Zhou Y. Tempering of martensite in dualphase steels and its effects on softening behavior. Metall Mater Trans A 2011;42(11):3116–28. [16] Nayak SS, Baltazar Hernandez VH, Zhou Y. Effect of chemistry on nonisothermal tempering and softening of dual-phase steels. Metall Mater Trans A 2011;42(11):3243–8. [17] Dancette S, Massardier-Jourdan V, Fabregue D, Merlin J, Dupuy T, Haz Bouzekri M. Microstructures and local mechanical properties of high strength steels resistance spot welds. ISIJ Int 2011;51(1):102–6. [18] Sun X, Stephens EV, Khaleel MA. Effects of fusion zone size and failure mode on peak load and energy absorption of advanced high strength steel spot welds under lap shear loading conditions. Eng Fail Anal 2008;15(4):359–63. [19] Khan MI, Kuntz ML, Zhou Y. Effects of weld microstructure on static and impact performance of resistance spot welded joints in advanced high strength steels. Sci Tech Weld Join 2008;13(3):295–303. [20] Baltazar Hernandez VH, Kuntz ML, Khan MI, Zhou Y. Influence of microstructure and weld size on the mechanical behaviour of dissimilar AHSS resistance spot welds. Sci Tech Weld Join 2008;13(8):772–6. [21] Baltazar Hernandez VH, Panda SK, Okita Y, Zhou NY. A study on heat affected zone softening in resistance spot welded dual phase steel by nanoindentation. J Mater Sci 2010;46(6):1640–6. [22] Baltazar Hernandez VH, Panda SK, Kuntz ML, Zhou Y. Nanoindentation and microstructure analysis of resistance spot welded dual phase steel. Mater Lett 2011;64(2):208–10.

609

[23] Choi HS, Park GH, Lim WS, Kim BM. Evaluation of weldability for resistance spot welded single-lap joint between GA780DP and hot-stamped 22MnB5 steel sheets. J Mech Sci Technol 2011;25(6):1543–50. [24] Kim CH, Choi JK, Kang MJ, Park YD. A study on the CO2 laser welding characteristics of high strength steel up to 1500 MPa for automotive application. J Achievements Mater Manuf Eng 2010;39(1):83–4. [25] Hu1 L, Yan Q, Huang J, Wu Y. microstructures and mechanical properties on laser welded joints of automotive high strength complex phase steel. Adv Mater Res 2010;97(101):3958–60. [26] Park TJ, Yu JW, Kang JI, Han TK, Chin KK, Kang CY. Effect of B contents on hardness characteristic of disk laser beam welded CP steels. J Korean Weld Soc 2011;29(1):109–13. [27] AWS/SAE D8.9. Recommended practices for test methods for evaluating the resistance spot welding behavior of automotive sheet steel materials. An American National Standard; 2002. [28] JIS Z 3136. Specimens dimensions and procedure for shear testing resistance spot and embossed projection welded joints. Japanese Standards Association: Tokyo (Japan); 1999. [29] Pouranvari M, Asgari HR, Mosavizadch SM, Marashi PH, Goodarzi M. Effect of weld nugget size on overload failure mode of resistance spot welds. Sci Tech Weld Join 2007;12:218. [30] ASTM F1529. Standard test method for sheet resistance uniformity evaluation by in-line four-point probe with the dual-configuration procedure. American Society for Testing and Materials; 1997. [31] Ishikawa N, Shikanai N, Kondo J. Development of ultra-high strength linepipes with dual-phase microstructure for high strain application. JFE Tech Rep 2008;12:16–7. [32] Heuschkel J. The expression of spot-weld properties. Weld J 1952;31(10):939–40. [33] Kabasawa M, Funakawa Y, Ogawa K, Tamura M. Estimation of tensile shear strength of spot welded joint of steel sheets. J Jpn Weld Soc 1996;14:754–8. [34] Khan MI, Kuntz ML, Biro E, Zhou Y. Microstructure and Mechanical Properties of Resistance Spot Welded Advanced High Strength Steels. Mater Trans 2008;49(7):1630–6. [35] Ma C, Chena DL, Bhole SD, Boudreau G, Lee A, Biro E. Microstructure and fracture characteristics of spot-welded DP600 steel. Mater Sci Eng A 2008;485:338. [36] George E. Dieter. Mechanical metallurgy. SI metric ed. New York: McGraw-Hill Inc.; 1988. p. 79. [37] ASTM E112-96. Standard test methods for determining average grain size; 2004.