1 Introduction

Dual-phase (DP) steel primarily consists of hard martensite embedded in a soft ferrite matrix. It offers an optimal combination of strength–ductility, continuous yielding and moderately high work hardening rate, making it lightweight—high-strength material (Ref 1,2,3). In general, DP steels cover grades with tensile strengths ranging from 490 to 1180 MPa, and specifically, DP 600 and DP 800 steels are commonly used for front side members, A, B, C pillars, cross members and the roof tail portion of a car body (Ref 3). In automotive industries, laser welding technology is proven to be more effective and convenient than other high energy density beam welding processes for thinner sections since it does not require vacuum, reduces the loss of alloying elements and can be easily automated by utilizing industrial robots (Ref 4). It also has the advantages over conventional fusion welding processes, such as low heat input, high efficiency, minimal distortion in welds and laser beam flexibility making the process highly productive and aesthetically attractive. Depending on the laser welding process conditions, a continuous or pulsed mode is employed. A laser emits a steady-state beam over time in continuous mode, while controlled short pulses are emitted in pulse wave mode (Ref 5). Based on Scopus indexed literatures, Fig. 1 summarizes the number of published articles on laser welding of DP steel over a period more than two decades, illustrating the exponential growth of research in this field. It is found that majority of the previous studies were based on continuous mode of laser welding and these studies were focused on the effect of heat input on the evolution of microstructures, heat-affected zone (HAZ) softening and mechanical properties (Ref 6,7,8,9,10,11,12,13). Previous researches showed the formation of martensite in the fusion zone (FZ), whereas tempering of existing martensite was observed in the HAZ. In DP steels, softening of HAZ in welded joints (WJ) easily occurs, resulting in a decrease in hardness and strength of softened zone compared to the base material (BM) and thus bearing significant effect on the mechanical properties of WJ.

Fig. 1
figure 1

The number of Scopus-indexed papers on “Laser welding of dual-phase (DP) steels” from 1998 to 2022

Full size image

Xia et al. (Ref 6) studied the HAZ softening of DP steels of various grades having tensile strength of 450 to 980 MPa and found that softening of HAZ is directly proportional to both the martensite content and heat input. More HAZ softening occurred with higher martensite content and heat input. Wang et al. (Ref 7) identified two softening mechanisms in laser-welded DP 1000 steel joints-tempering of martensite in softened HAZ with consequent precipitation of carbides, and a reduction in martensite content in the intercritical heat-affected zone (ICHAZ) due to the formation of softer phases such as ferrite, retained austenite (RA) and bainite during cooling. According to Alves et al. (Ref 8), both presence of RA and tempering of prior martensite are responsible for softening of HAZ of DP 1000 steel laser WJ. In contrast, recently, Saha et al. (Ref 9) showed that the presence of blocky RA in steel improved WJ properties with 100% joint efficiency and this happened due to transformation of untempered RA into strain-hardened martensite. Jia et al. (Ref 10) investigated the effect of heat input on DP 980 steel laser WJ using both experimental and numerical methods and found that tensile strength of joints decreased with increase of heat input. The study also revealed an increase in HAZ width with increase of heat input, resulting in fracture in the softened HAZ under tensile loading with low fracture strain, while the WJ with lower heat input failed in the BM because of the biaxial stress concentration effect. Wang et al. (Ref 11) examined the effect of heat input on the microstructure and fracture behavior of DP 780 steel WJ and found that with increase of heat input tensile strength, elongation and formability of WJ decreased. The study also revealed an increase in martensite decomposition and carbide precipitation in the softened HAZ as the heat input increased, resulting in a decrease in the WJ mechanical properties (Ref 11). Chen et al. (Ref 12) studied the effect of heat input on the microstructure and mechanical properties of 800 MPa high-strength low-alloy steel WJ and found that as heat input decreased, the FZ microstructure changed from acicular ferrite to martensite. The width of HAZ and prior austenite grain size in the HAZ region increased with increase of heat input. It was also found that increased joint strength with low heat input was due to the formation of martensite in the FZ and narrow HAZ (Ref 12). Wang et al. (Ref 13) also observed a decrease in tensile strength and elongation of DP 1000 steel WJ with an increase in heat input due to the formation of extended softened HAZ region.

Pulsed laser welding, on the other hand, has many advantages over continuous-mode laser welding, including lower average power, low heat input, higher cooling rate, reduced FZ and HAZ widths, low residual stress and reduced HAZ softening (Ref 3, 14,15,16). Though the average power of pulsed wave lasers is less, the peak power at each pulse is high, resulting in a high effective peak power density that helps in the creation of deep keyholes and high aspect ratio profiles (Ref 17, 18). As a result, it has significant advantage in thin sheet welding, particularly dissimilar materials that are commonly used in tailor-welded blanks in automotive industry (Ref 19). Continuous mode laser welding of galvanized steel creates weld defects such as blowholes, spatter and poor weld surface quality (Ref 20). Pulsed mode, on the other hand, enables pulse shaping, which controls the cooling rate of the weld pool and thus facilitates the joining of galvanized steel sheet and fabrication of defect-free joints (Ref 3). Thus, the utilization of low-power lasers offers the potential to reduce both the initial cost of production and energy consumption.

Despite several advantages, research work on pulsed laser welding of DP steel joints is very scanty (Ref 3, 14, 21, 22). The present authors compared continuous and pulsed laser welding in some detail and found that higher pulse duration improved the WJ mechanical performance due to reduced WJ concavity (Ref 3). The study also showed that with an increase in pulse duration, the effective peak power density decreased, simulating the low heat input effect and resulting in a reduction in WJ bead size. Sun et al. (Ref 14) investigated the effect of pulse frequency on laser-welded DP 590 steel joints and found that as the frequency increased, the overlapping factor improved, replicating an increase in average power and heat input, which led to a deeper weld. Additionally, microstructure of FZ changed from lath martensite to bainite and ferrite. The study also showed improvement of tensile properties and formability with an increase of pulse frequency. Xue et al. (Ref 21) also reported that higher the pulse duration and welding speed, the deeper was the weld penetration for pulsed laser-welded DP 1000 steel joints. In addition, Xue et al. (Ref 21) observed lath martensite along with ferrite and bainite in the FZ. Tuncel et al. (Ref 22) also reported that tensile properties, e.g., tensile strength, total elongation of pulsed laser-welded DP 600 steel joints improved almost linearly with an increase of pulse frequency. However, to the authors’ best knowledge no report is available on the effect of heat input on pulsed laser-welded DP steel joints in the open literature access.

The above review reveals that there is complete dearth of knowledge on the effect of laser pulse on weld thermal cycles by taking solidification latent heat into account in pulsed laser welding of DP steels. Systematic investigation of pulsation effect on microstructural evolution, tensile properties and fracture behavior in pulsed laser-welded DP steel joints is also lacking. The present work is a step towards addressing the aforesaid issues. In the present investigation, pulsed laser welding of DP 780 steel in a butt joint configuration was performed with varying heat input from 32.5 to 92.22 J/mm. The effect of pulsation phenomenon on the weld thermal cycle was studied using a numerical approach, and corresponding microstructural evolution was analyzed. Further, microhardness variation and tensile behavior of WJs were also investigated. The correlation of the weld thermal cycles with the evolution of microstructures in the FZ, HAZ on the tensile deformation and fracture behavior has been thoroughly examined.

2 Experimental Procedure

Dual-phase steel of grade DP 780 was received in sheet form with a nominal thickness of 1.00 mm from TATA Steel Limited, Jamshedpur. The chemical composition of DP 780 steel is given in Table 1. The Thermo-Calc software was used to calculate the liquidus and solidus temperatures and the temperatures of phase transformations of DP 780 steel. The solidification start and finish temperatures of DP 780 steel, estimated using Thermo-Calc, were 1790 and 1748 K, respectively (Ref 3). The critical transformation temperatures Ac3 and Ac1 were calculated as 1096 and 964 K, respectively.

Table 1 Chemical composition of DP 780 steel
Full size table

The pulsed laser welding was performed in a JK 600 HP with a Nd:YAG laser generator (GSI, UK) integrated with an ABB IRB 1410 robotic control; the welding setup is described in our earlier report (Ref 3). Welding was performed perpendicular to the rolling direction of the sheet in butt joint configuration (Fig. 2a) using the parameters listed in Table 2. A three-dimensional (3-D) finite element (FE) model was developed using ABAQUS 2017 software to simulate the pulsed laser welding process. In the FE model, proper considerations of the heat dissipation phenomenon are crucial because of their great significance on the temperature field, shape and size of the laser weld bead. For meshing, the numerical model employs 3D solid heat transfer elements of the DC3D8 type. Fine mesh (element size 0.05 × 0.05 mm) is used for FZ and HAZ to capture the steep temperature gradient, while the rest of the model uses a progressive coarse mesh (element size 0.05 × 0.10 mm). The spatial and temporal variations, including the pulse on and off of the moving heat source, are described using the DEFLUX subroutine. The model uses an hourglass shape heat source with a Gaussian distribution because of its ability to map deep penetration (Ref 3). The temperature-dependent thermo-physical properties, assumptions and boundary conditions applied to the model are well described in the previous work (Ref 3).

Fig. 2
figure 2

(a) Schematic representation of the specimen used for laser welding, (b) position of microhardness indentation taken on the cross section of weld microstructure and (c) geometry of the laser-welded tensile specimen as per ASTM E8M standard

Full size image
Table 2 Parameters used for laser welding
Full size table

The transverse cross section of welds was machined and mounted on a Struers CitoPress-5 using bakelite. It was then polished in an automatic polishing machine Struers TegraPol 21. To reveal the microstructural constituents present in FZ, HAZ and the BM, the weld specimens were etched with 2% Nital and Lepera etchant (Ref 8). Microstructure was examined in an optical microscope, Leica DM 2700 M coupled with image analysis software, LAS V4.4, as well as field emission scanning electron microscope (FEGSEM), FEI Quanta 450. The Vickers microhardness was measured on the weld specimens (Fig. 2b) in a MATSUZAWA, Japan make automatic microhardness tester with 100 g load and 15 s dwell time. To prevent interference between the successive indentations, a distance of approximately 100 µm was maintained between them.

Cross-weld tensile specimens (Fig. 2a and c) were prepared from welded joints keeping the long axis of tensile specimens parallel to rolling direction of the sheet. The specimens were (Fig. 2c) fabricated as per the ASTM E8M standard. Strain-control tensile tests up to fracture were performed at a strain rate of 1 × 10−3 s−1 in a computer-controlled servohydraulic universal testing machine, INSTRON 8501R of ± 100 kN capacity. A 25.00 mm gauge length extensometer with a full range capacity of + 50 to −10% was used for strain measurement. The data were collected using Instron Bluehill tensile testing software. Fractured surfaces of tensile specimens were observed in the same FEGSEM that has been used for studying microstructure. The image analysis technique was used to characterize the fractured tensile specimens in terms of dimple size and their distribution in BM and WJ. At least two hundred fifty dimples were identified in the fibrous zone and analyzed from multiple fractographs to cover the entire fracture surface. The failed tensile specimens were sectioned perpendicular to the fracture surface through the transverse plane. It was then metallographically prepared to study the microscopic damage mechanisms and void initiating sites during tensile deformation of WJ.

3 Results and Discussion

3.1 Effect of Heat Input on Transient Thermal Profile

The finite element simulation was performed to understand the temperature profile and heat transfer phenomenon in pulsed laser welding. Figure 3 shows the isometric view of numerically simulated maximum temperature distribution along scan line during welding at four different time intervals as suggested in the literature (Ref 23). With a heat input of 32.5 J/mm, the peak temperatures of 1891, 1898, 1917 and 1955 K were obtained from the simulation at various time intervals (0.0013, 0.193, 0.35 and 0.573 s) showing the progressive increase of peak temperature with increase of time interval. Additionally, with increase of heat input from 32.5 to 72.22 J/mm peak temperature increases because it extends the interaction time between the moving heat source and the sheet surface during welding. As shown in Fig. 3, the temperature contours are expanded behind the laser beam and compressed in front of the beam due to the relative motion of the weld pool and the laser beam (Ref 24). Furthermore, as the welding progresses, the temperature contours expand more pronouncedly behind the weld pool due to the preheating experienced in successive pulses (Ref 24). In pulsed laser welding, energy applied to an area comes from both single and overlapping pulses on the same spot area (Ref 25). Hence, in order to characterize this feature of pulsed laser welding technique the overlapping factor is introduced to consider the contribution of the set of pulses that hit on the same area of spot (Ref 25). Overlapping factor indicates the energy required for melting of a material (Ref 14). With an increase in overlapping factor, both heat input per unit length and melting rate are increased (Ref 26). Generally, the melting rate defined as the ratio of energy required to melt the weld metal to the energy supplied to BM is strongly dependent on the pulse overlap and welding speed (Ref 27). The overlapping factor is calculated from Eq 1 (Ref 3, 14) and listed in Table 2.

$$O_{f} = \left( {1 - \frac{{\left( \frac{V}{f} \right)}}{{\left( {D + V*t_{p} } \right)}}} \right)*100$$
(1)

where V is welding speed, f is pulse frequency, tp is pulse duration, and D refers to a spot diameter of the laser beam.

Fig. 3
figure 3

The transient temperature distribution on the top plane of the sheet for the welding conditions: (a) heat input: 72.22 J/mm and (b) heat input: 32.5 J/mm at different time intervals

Full size image

As listed in Table. 2, there is no significant difference between the overlapping factor for 72.22 and 92.22 J/mm when compared to 32.50 and 72.22 J/mm heat input welding conditions. In pulsed laser welding, each pulse has its own thermal cycle, and a continuous seam weld results from the overlapping of individual successive pulses. Therefore, the microstructure and mechanical properties of WJs are strongly influenced by the overlapping factor during pulsed Nd:YAG laser welding (Ref 25). Sheikhbahaee et al. (Ref 28) have shown that increasing the power has a relatively small effect on spot overlapping in the low speed region, which is consistent with the findings of the present study at 72.22 and 92.22 J/mm heat input conditions (Table 2). Zhao et al. (Ref 29) also observed that welding speed, which significantly affected the weld thermal cycle, was the most influential factor to control the microstructure, hardness, tensile properties of laser-welded thin gauge steel. Guided by the aforesaid knowledge we have chosen the 32.50 and 72.22 J/mm heat input conditions to study the influence of heat input on pulsed laser WJ in detail. However, the transient temperature distribution for the heat input of 56.25 and 92.22 J/mm is also provided in Appendix, Fig. 18. Validation of pulsed laser numerical model and its pulsation effect are difficult to conduct with conventional experiments (Ref 30). So, the compatibility of a model needs to be verified by comparing the numerically computed weld pool dimensions at the mid-thickness with the experimental results for different heat input conditions.

Figure 4(a) and (b) compares the numerically computed and experimentally obtained weld pool dimensions. The FZ and softened HAZ are defined by the isotherms. The red contour represents the FZ with temperatures ranging from 1964 to 2116 K for 72.22 J/mm and 1759 to 1891 K for 32.5 J/mm heat input WJs. Higher heat input results in larger weld bead size and higher peak temperature during welding. The softened HAZ is shown by a light blue and green color band with a temperature range of 752-1055 K for 72.22 J/mm heat input WJ (Fig. 4a) and a light green color band with a temperature range of 696-829 K for 32.5 J/mm heat input WJ (Fig. 4b). The softened HAZ involves solid–solid phase transformation. Wang et al. (Ref 7) identified that depending on the heat input HAZ softening occurred in temperature ranges from 623 to 1166 K for DP steel welded joints. The higher the heat input, the higher is the temperature in the softened HAZ zone resulting in more HAZ softening (Ref 6). Accordingly, the zones are identified and located from weld centerline with the help of experimental bead size values. Additionally, the location of softened HAZ region from the weld centerline is also confirmed by the microhardness indentation technique which is discussed in the later microhardness profile section. In the present simulation, the numerically computed and experimentally measured FZ sizes at mid-thickness of WJs are in good agreement, with a maximum error of 8.2%. Similar, validation of the model for heat inputs 56.25 and 92.22 J/mm is provided in Appendix, Fig. 19. The choice of proper boundary conditions to incorporate the heat losses caused by different modes of heat transfer is crucial for accurate simulation. Therefore, the surfaces of steel sheets which are in direct contact with air are assigned with convective heat transfer coefficient of air as 25 W/m2 K (Ref 31). However, the bottom surface of steel sheet is in close contact with the thermally conductive copper backing plate. In order to compensate for this thermal contact resistance Baruah and Bag (Ref 30) suggested to assign a higher heat transfer coefficient to the bottom of the surface of the steel sheet. In this regard, it is suggested to consider the heat transfer coefficient ranging from 500 to 2000 W/m2 K for representing the impact of the backing plate on heat transfer by convection (Ref 32). Within this range, different values of convective heat transfer coefficient were investigated, and a good agreement between the numerical value and experimental data was observed at 1000 W/m2K, which was then used to the bottom surface of the steel sheet that was in contact with copper backing plates. The latent heat of fusion for the present steel is 240,000 J/kg (Ref 33). Radiation heat transfer with an emissivity value of 0.34 was used to the top surface of the plate (Ref 33). The time step considered for simulation is of at least 2.5 times the welding time in order to account for the cooling of steel sheets.

Fig. 4
figure 4

Comparison of FE simulations and experimental results for welding conditions with different heat inputs: (a) 72.22 J/mm and (b) 32.5 J/mm

Full size image

Figure 5(a) and (b) depicts a nodal time-temperature profile for various conditions that can estimate weld pool temperature and cooling rate responsible for weld solidification structure. It demonstrates the pulsation effect caused by fluctuations in the temperature profile. In case of pulsed laser welding, BM is exposed intermittently to a high energy density laser beam for a short duration, resulting in rapid melting and solidification in a small spot area. The temperature rises sharply during the pulse on time (Fig. 5a and b) to ensure melting and drops down to a lower value. It is of worth mentioning that each thermal cycle is different from the previous thermal cycles, which causes the temperature to rise due to the preheating of BM in successive pulses. Therefore, at any location (X) away from the weld centerline (Fig. 5a and b), BM is subjected to several thermal cycles during the pulsed welding involving melting cycles that increase with the heat input. The weld thermal cycles include a periodic rise in temperature during the heating cycle and a rapid decrease in temperature until the liquidus temperature, followed by slope changes that appear as a nearly flat step on the cooling cycles, indicated by the red arrow in Fig. 6(a) and (b). These kinks in weld thermal cycles are caused by the latent heat changes associated with phase transformations that occur during the solidification of welds (Ref 34). It is also worth mentioning that the weld thermal cycle is unique for each location, and it determines the final solidification structure based on cooling rates calculated at critical temperatures associated with liquid to solid and solid–solid phase transformations. It is to be noted that because of smaller bead size, large temperature gradient and extremely high cooling rate, accurate measurements of FZ temperature and cooling rates are extremely difficult in laser WJ. Further, the equation used by Saha et al. (Ref 35) to calculate the cooling rate cannot be directly used to the pulsed laser WJ since this equation is based on the Rosenthal equation, which makes assumptions of using a point source to represent the welding arc, ignoring latent heat changes and convective heat losses. It also assumed that material properties remain unaltered by temperature changes and, most importantly, ignored the intermittent laser beam interaction involved in pulsed laser. To overcome these limitations, Zacharia et al. (Ref 36) calculated cooling rates based on the slopes of weld thermal cycles obtained from numerical simulations. The same technique has been used in the current study, and cooling rates have been computed for weld thermal cycles with a peak temperature lower than the melting temperature (Tm). For high heat input (72.22 J/mm) and low heat input (32.5 J/mm) WJs, the cooling rates have been calculated at different times, e.g., t1, t2, t3 etc. (Fig. 6), from the changes in slope at the same location away from the weld centerline (X). The slopes are identified by the rate of change of temperature (dT/dt) (Fig. 6a and b). The cooling rates, estimated in the FZ of high input WJ (72.22 J/mm) at location X = 0.29 mm at different times intervals (0.51, 1.24, 1.34, 1.67, 2.41 and 2.78 s) are 6882, 416, 935, 954, 547 and 250 K/s, respectively (Fig. 6a). However, for low heat input WJ (32.22 J/mm), cooling rates, calculated at a location X = 0.12 mm at different times (0.30, 0.35, 0.39, 0.41, 0.44, 0.48, 0.54 and 0.75 s), are 16743, 13509, 8552, 3251, 3251, 3898, 1763 and 933 K/s, respectively (Fig. 6b). As time progresses, the same location (X) is exposed to different temperatures as schematically shown in Fig. 6(a) and (b), thus resulting in different temperature gradients. The FZ cooling rate is determined by the temperature gradient and the capacity of the adjacent base metal to remove heat from the fusion zone (Ref 36), whereas in case of a continuous laser, the cooling rate decays exponentially over time (Ref 37); the pulsed laser, on the other hand, produces several temperature peaks in the weld thermal cycle providing additional heat to the materials. As a consequence, alteration of the current temperature gradient occurs through change of the location and thereby the cooling rate (Ref 36). So, the observed solidification microstructure as discussed in the next section is the cumulative effect of multiple complex laser-metal interactions. It is worthwhile to mention that the pulsation effect that determines the temporal and spatial variation of the temperature profile, resulting in non-uniformities in cooling rates, can significantly modify the solidification microstructure and thus properties of WJ.

Fig. 5
figure 5

Calculated temperature profile given at any location X from the weld centerline and taken at the center of a sheet for different heat inputs: (a) 72.22 J/mm and (b) 32.5 J/mm, with a zoomed-inset view of the pulsation effect

Full size image
Fig. 6
figure 6

Calculated weld thermal cycle experienced in FZ at a distance (X) from the weld centerline, located at the center sheet for different heat inputs: (a) 72.22 J/mm and (b) 32.5 J/mm and the schematic representation of the simulated weld pool temperature contours at the different time intervals marked as (t1), (t2), (t3), etc.

Full size image

3.2 Microstructure of the Welds

Microstructures of various weld cross-sectional regions for all the heat inputs are shown in Fig. 7, 8, 21 and 22, and Table 3 presents phase fraction of the WJs. Weld cross section is generally categorized into FZ, HAZ and BM. The weld cross-sectional macroview had an inverted conical shape with an average FZ width of 0.33 ± 0.20 mm at the lowest heat input of 32.5 J/mm and a slightly narrower rectangle cross section with an average FZ width of 0.63 ± 0.38 mm for the highest heat input of 92.22 J/mm. With increase of heat input, the average width of FZ and softened HAZ (marked in macro views) increases continuously. Table 3 also shows the detailed weld bead dimensions. As shown in Fig. 7 (with a magnified view in the inset), microstructure of DP 780 BM contains ferrite and martensite phases. The amount of martensite was 32 ± 2% by area. In the present study, HAZ can be divided into subcritical HAZ, also considered as softened HAZ (SHAZ), while intercritical HAZ, fine grain HAZ (FGHAZ) and coarse grain HAZ (CGHAZ) are regarded as hardened HAZ.

Fig. 7
figure 7

Microstructures of various regions of the weld cross section at a heat input of 72.22 J/mm, including the base material

Full size image
Fig. 8
figure 8

Microstructures of various regions of the weld cross section at a heat input of 32.5 J/mm

Full size image
Table 3 Details of microstructural constituents in the fusion zone of the welded joints
Full size table

Figure 7, 8, 9, 21 and 22 shows that the FZ microstructure predominantly consists of lath martensite (LM), along with ferrite (F) and bainite (B) for all the heat inputs. Mostann et al. (Ref 38) also found the lath martensite and bainite microstructure in the FZ of pulsed laser-welded DP steel joints. In the present study Lepera tint etching (Ref 8) was used to delineate martensite by white color and ferrite-bainite constituent by tan color (Fig. 9). The amount of ferrite-bainite content was measured by using the color threshold technique as shown in Fig. 9(e) and (f), and the amount of phases is listed in Table 3. It is found that the amount of ferrite-bainite content increases as the heat input increases and this happens due to higher peak temperatures and lower cooling rates associated with high heat input (Fig. 3 and 5). The peak temperature estimated by the numerical FE model in this zone was much higher than the liquidus temperature for both heat inputs of 32.5 and 72.22 J/mm (Fig. 4a and b), which ensures complete melting. As already mentioned, the high cooling rate of the order 102 to 104 K/s experienced during weld solidification results in a predominantly lath martensite microstructure. However, in the case of a pulsed laser, the region near the solidified front is reheated to nearly the solidus temperature during subsequent pulses. As a result, it is likely that in this region solid–solid phase transformation (austenite to δ-ferrite) (Ref 36) would occur. The delta ferrite (δ-ferrite) though stable only at high temperatures and small amount of this phase is likely to remain in the structure because of very high cooling rate of the order of 104 K/s experienced during cooling. Such high cooling rate is sufficient to suppress the transformation of δ-ferrite to austenite and thus to martensite. Hence, more ferrite content, observed in the FZ, in case of high heat input WJ (Fig. 7) is presumably due to more pronounced pulsation effect during welding. Saha et al. (Ref 39) also observed the FZ microstructure containing lath martensite and non-equilibrium high-temperature δ-ferrite phase in laser-welded press hardened steel joints. The FEGSEM microstructures of FZ show the formation of lath martensite (LM) along with coalesced bainite (CB) and lower bainite (LB) for both heat inputs of 32.5 and 72.22 J/mm (Fig. 7 and 8). Correard et al. (Ref 40) also observed the formation of coalesced bainite and lath martensite in the FZ of laser-welded DP 600 steel joints. Coalesced bainite occurs when many supersaturated bainitic ferrite plates begin to form and coalesce into one big grain during growth (Ref 41). Formation of large prior austenite grains at higher heat input (Table 3) is conducive in the reconstructive transformation generating more ferrite and coalescence process, which ultimately leads to more coalesced bainite formation (Ref 42).

Fig. 9
figure 9

Lepera etched optical microstructures of FZ for identifying microstructural constituents at various heat inputs, including (a) BM, (b) 32.5 J/mm, (c) 56.25 J/mm, (d) 72.22 J/mm, (e) 92.22 J/mm and (f) a color thresholding image of 92.22 J/mm used for illustrating phase fraction calculation

Full size image

The CGHAZ for both heat inputs (Fig. 7 and 8) shows transformed lath martensite and coalesced bainite microstructure within reaustenitized grains during welding. The peak temperature in this zone is slightly lower than in FZ; hence, heat dissipation rate becomes low in this zone in comparison with the other HAZ zones resulting in rapid austenite grain growth in this zone. The prior austenite grain size was measured and found as 9.4 ± 2.40, 16.20 ± 3.40, 19.2 ± 3.60 and 20.3 ± 4.05 µm for 32.5, 56.25, 72.22 and 92.22 J/mm heat input welds, respectively. In FGHAZ (Fig. 7 and 8), lath martensite and fine grain ferrite (FGF) were found within prior austenite grains that were smaller in size than in CGHAZ. For 32.5, 56.25, 72.22 and 92.22 J/mm heat input welds, the prior austenite grain size in the FGHAZ region was 4.83 ± 1.30, 5.35 ± 1.5, 7.06 ± 1.17 and 7.45 ± 1.20 µm, respectively. The FGHAZ region is away from the weld centerline and experiences lower peak temperature resulting in smaller prior austenite grains (Ref 43). Microstructure of ICHAZ (Fig. 7 and 8) is comprised of martensite (M) and ferrite (F). Compared to higher heat input weld (72.22 J/mm), lower heat input weld (32.5 J/mm) had large blocky martensite in this zone (Fig. 7 and 8).

In both WJs, the softened HAZ (SHAZ) is composed of tempered martensite (TM) and ferrite (F). The estimated peak temperature in this zone is in the range of 696-829 K and 752-1055 K for low (32.5 J/mm) and high heat input (72.22 J/mm) welds. Tempered martensite is observed in this zone due to the decomposition of pre-existing martensite in the BM, including recovery or recrystallization of martensite and precipitation of carbides (Ref 11). Recovery occurs during tempering of martensite at temperatures ranging from 623 to 873 K, with significant rearrangement of dislocations within the martensite laths and at the lath boundaries (Ref 3). This recovery process is replaced by recrystallization in the temperature range of 873 to 973 K, resulting in the formation of equiaxed ferrite grains indicated by the yellow arrow and spheroidal cementite particles at the grain boundaries and also within the grains indicated by blue arrow in Fig. 10(a) and (b).

Fig. 10
figure 10

Effect of heat input on the microstructure of softened HAZ, showing the formation of tempered martensite (TM) and ferrite (F) for the heat inputs of (a) 72.22 J/mm and (b) 32.5 J/mm, with the tempering of martensite resulting in the formation of ferrite grains shown by the yellow arrow and cementite precipitation indicated by the blue arrow at and within grain boundaries (Color figure online)

Full size image

In general, tempering of martensite depends on heating temperature and holding time at that temperatures. Due to the non-isothermal tempering process experienced during laser welding, the time required to heat the BM to a high temperature is the key factor in the formation of tempered martensite (Ref 11). Xia et al. (Ref 6) calculated a time constant (τ) for comparing tempering times of two welds produced by Nd:YAG and diode laser processes, which represents the time required to heat the BM to its Ac1 temperature, given by Eq 2

$$\tau = \frac{1}{4\pi e\lambda \rho c}.\frac{{\left[ {\frac{{Q_{{{\text{net}}}} }}{V.g}} \right]^{2} }}{{\left( {T_{p} - T_{0} } \right)^{2} }}$$
(2)

where Qnet is net laser power (W), V is the welding speed (m/s), g is the sheet thickness (m), ρ is the density of steel (7860 kg/m3), λ is the thermal conductivity (30 W/m k), Tp and To is the peak temperature at any location and ambient temperature, respectively (Ref 6). Following Eq 2, the time constants (τ) for 72.22 and 32.5 J/mm heat input WJs are estimated as 1.67 and 0.68 s, respectively. Figure 11(a), (b), (c) and (d) shows Lepera-etched HAZ microstructures with varying heat inputs. At lower heat input of 32.5 J/mm (Fig. 11a), martensite remains as white etching contrast, indicating minimal submicroscopic carbides. In contrast, at higher heat input of 72.22 J/mm (Fig. 11c), martensite becomes more patchy and darker, indicating the formation of significant submicroscopic carbide due to severe tempering (Ref 6). With an increase in heat input from 32.5 to 92.22 J/mm, the fraction of the second phase (tempered martensite, carbides) and ferrite grain size increased (Fig. 11e) due to a higher extent of tempering in the softened HAZ.

Fig. 11
figure 11

Effect of heat input on HAZ softening, illustrating the extent of martensite tempering by Lepera-etched microstructures for the heat inputs of (a) 32.5 J/mm, (b) 56.25 J/mm, (c) 72.22 J/mm and (d) 92.22 J/mm and (e) variation of phase fractions and ferrite grain size within the softened HAZ with respect to different heat inputs

Full size image

3.3 Microhardness Profile

Figure 12(a) shows the Vickers microhardness profile of the WJs prepared with different heat inputs. The weld with a heat input of 32.5 J/mm is considered to explain the microhardness variation in each zone of a joint. The average hardness value of BM was 250 HV0.1, whereas FZ shows significantly higher hardness of 413 HV0.1, consistent with the presence of predominantly lath martensitic microstructure (Fig. 8 and Table 3). Due to tempering of prior martensite, hardness of the outer HAZ near the BM is lower than the BM and this outer HAZ is termed as softened HAZ. The hardness value of this zone lies within the range of 244-248 HV0.1 for lower heat input of 32.5 J/mm and 221-234 HV0.1 for higher heat input of 72.22 J/mm WJs. In the case of 56.25 and 92.22 J/mm heat input, hardness of softened HAZ is found as 232-236 HV0.1 and 218-232 HV0.1, respectively. Further from the microhardness profile, the location of the softened HAZ is found to depend on the heat input. While with a heat input of 32.5 J/mm the formation of softened HAZ occurred at 0.45-0.60 mm away from the weld centerline, the same was observed at 0.92-1.11 mm away from the weld centerline for the heat input of 72.22 J/mm. Also, according to FE simulation the peak temperatures experienced in these softened HAZ regions are in the range of 696-829 to 752-1055 K (Fig. 4a and b) for low (32.5 J/mm) and high heat input (72.22 J/mm) WJs. It is thus found that the rise in temperature depending on the heat input is the governing factor on the variation in microhardness results. Severe tempering with a consequent drop in hardness in the softened HAZ occurs as the rise of temperature becomes closer to AC1 temperature (Ref 7) as observed for welding with a higher heat input of 72.22 J/mm.

Fig. 12
figure 12

(a) Microhardness profile of WJs with the different heat inputs and (b) average microhardness variation in the FZ and HAZ of WJs as a function of the heat input

Full size image

Figure 12(b) compares the average hardness variation with heat input in different regions of WJs. With an increase in heat input from 32.50 to 92.22 J/mm, the presence of higher amount of δ-ferrite within the predominant martensitic microstructure in FZ promotes local softening, thus causing a reduction in hardness from 399 to 358 HV0.1. Prior austenite grain (PAG) size which is found to increase with an increase in heat input has a strong effect on morphology of the phases that forms on cooling. Since the acicularity of martensite for a given steel composition depends on PAG size in such a manner that with lowering of PAG size the acicularity of martensite increases, the PAG size has a significant effect on the strengthening of fine-grained microstructure. As a consequence, hardness of hardened HAZ in the present study is found to increase from 287 to 312 HV0.1 with a decrease of heat input from 92.22 to 32.5 J/mm. On the other hand, the increase of the average hardness of softened HAZ from 230 to 244 HV0.1 with decrease of heat input from 92.22 to 32.5 J/mm is due to the decrease in the extent of tempering in this zone.

3.4 Effect of Heat input on Tensile Properties, Fracture and Deformed Microstructure of Welds

Figure 13 depicts typical engineering stress–strain curves for both BM and WJs. All stress–strain curves show continuous yielding behavior due to the presence of high mobile dislocations in the BM and because of the presence of residual stress in the WJs (Ref 44, 45). Table 4 shows the tensile test results. The efficiency of WJs is calculated as the ratio of WJ tensile strength to the corresponding BM tensile strength (Ref 46). The uniform elongation is calculated from the point of plastic instability criteria, where the work hardening rate intersects the true stress–strain curve. The joint efficiency for all WJ is found as 92-110%; however, the maximum uniform elongation was only 57% of the BM (Table 4). Lower heat input WJ has higher YS and TS than BM, while WJ with higher heat input has slightly lower YS and TS (Fig. 14 and Table 4). This is related with narrow soft zone width and small hardness drop in WJ with lower heat input. Lin et al. (Ref 47) and Baghjari et al. (Ref 48) reported that presence of about 30-35% δ-ferrite in the martensitic microstructure can lead to cracking in FZ during weld solidification. In the present study, the amount of retained δ-ferrite in FZ was ~ 10-15%; hence, no adverse effect of δ-ferrite was observed in the FZ.

Fig. 13
figure 13

Engineering stress–strain curves of DP 780 BM and welded joints with the different heat inputs

Full size image
Table 4 Tensile test results of DP 780 BM and laser-welded joints
Full size table
Fig. 14
figure 14

Effect of heat input on the tensile properties of welded joints like yield strength, ultimate tensile strength and uniform elongation

Full size image

The locations of failure under tensile loading for the WJs prepared with 32.5, 56.25, 72.22 and 92.22 J/mm heat inputs are shown in Fig. 15(a), (b), (c) and (d), respectively. It is observed that irrespective of heat input all the welded specimens fail in the softened HAZ. This is attributed to significant tempering effect of martensite in the softened HAZ vis-à-vis accumulation of plastic deformation in this zone during tensile deformation.

Fig. 15
figure 15

Effect of heat inputs on the tensile failure locations indicated by a red arrow of DP 780 laser-welded steel joints with heat input: (a) 32.5 J/mm, (b) 56.25 J/mm, (c) 72.5 J/mm and (d) 92.22 J/mm

Full size image

Figure 16 shows the formation of equiaxed dimples on the fracture surfaces of WJs prepared by different heat inputs. Smaller dimples up to 4 µm in size are observed in specimen with lower heat input (32.5 J/mm), whereas the bigger dimples of 6-8 µm in size constitute the fracture surface of higher heat input (72.22 J/mm) weld specimen (Fig. 16e). The deformed microstructures of WJ near the tensile fracture location (Fig. 17) were studied to understand the tensile failure mechanism. Similar kinds of deformed microstructures were observed for all the heat inputs. As a result, we considered 32.5 J/mm and 72.22 J/mm heat input conditions to study the effect of heat input on the tensile failure mechanisms of welded joints. The deformed microstructures (Fig. 17) indicate the formation of void sheets; void initiation occurring at inclusions and through ferrite-martensite interface decohesion. The formation of void sheets occurred for all the heat inputs (Fig. 17a, b, c and d). It occurs through nucleation, growth and coalescence of voids, leaving a zig-zag profile on the fracture surface with its segments oriented to some degrees to the principal stress axis (Ref 49). A void sheet failure mode is reported to occur in steels containing a very limited number of elongated MnS inclusion (Ref 49). It involves the formation of large elongated microvoids, which are then linked together on an inclined plane by the coalescence of small secondary microvoids nucleated at fine carbide particles and inclusions (Fig. 17e). In the present case of lower heat input weld specimen, formation of a void sheet is observed where the void distance is small and failure occurs by plastic collapsing of the bridging material between them (Ref 50). It is likely to be caused by the formation of local shear bands between the ferrite grains, which began and ended at neighboring martensite grains and cut the ferrite grains between them (Ref 51) (Fig. 17j).

Fig. 16
figure 16

Typical SEM images of the fracture surface of tensile-tested DP 780 steel WJs under various heat input conditions: (a) 32.5 J/mm, (b) 56.25 J/mm, (c) 72.22 J/mm and (d) 92.22 J/mm and (e) dimple size distributions of the fracture surface (yellow arrow indicates a coalescence of voids that led to crack propagation and ductile failure) (Color figure online)

Full size image
Fig. 17
figure 17

Deformed microstructures of the transverse cross section, perpendicular to the fracture surface ends (a) to (d) showing the macroviews of the fracture locations for all heat inputs of 32.5 J/mm, 56.25 J/mm, 72.22 J/mm and 92.22 J/mm, respectively. In addition, the microscopic damage mechanisms for 72.22 J/mm and 32.5 J/mm are depicted in (e) to (g) and (h) to (j), respectively.

Full size image

Tempered martensite cracking occurs in case of high heat input weld specimen, indicated by the red arrow (Fig. 17e and f), and acts as void nucleation sites. In the case of tempered martensitic structure large and evenly distributed elongated cementite particles at PAG boundaries shared by two adjacent martensite blocks provide sites for intergranular cracking (Ref 52). As stiff cementite precipitates are not plastically deformable, it cracks or debonds at the martensite-cementite interface. Tempering, in general, is very effective in increasing martensite plasticity. As a result, separate void initiation site is observed in heavily strained tempered martensite in the necked zone, marked by a white arrow, (Fig. 17g). However, large and blocky martensite islands, on the other hand, are more prone to cracking during tensile deformation (Ref 53). In low heat input WJ, spherical voids are observed at the ferrite-martensite interfaces when martensite particles are closely spaced, with only small ferrite grain in between (Fig. 17i). Because of the spherical shape of the voids, it is believed that the presence of locally high hydrostatic stress between the martensite particles initiates the void formation (Ref 51). Hence, it leads to conclude that the amount and distribution of tempered martensite alters the void initiation sites.

4 Conclusions

Experiments and FE simulation were used to evaluate the influence of heat input in order to estimate the temperature profile and to find the effect of heat input on microstructural evolution in pulsed laser-welded DP steel joints and corresponding mechanical properties. The following conclusions have been drawn from the present study:

  1. 1.

    The present numerical model of pulsed laser welding predicted pulsation behavior in the weld thermal cycle involving fluctuations in temperature profile (multiple temperature peaks) in contrast to the typical exponentially decaying temperature distribution as reported in continuous mode laser welding. The pulsation effect increases with heat input and diminishes with the distance from the weld centerline.

  2. 2.

    The FZ and softened HAZ width of WJs increases with an increase of heat input. Experimental results and numerically estimated weld dimensions correlate satisfactorily, validating the numerical model of pulsed laser welding. The numerical model used in this study can effectively predict the size of the soft zone and the temperature range experienced in this region, providing guidance for welding DP steels.

  3. 3.

    Microstructure of FZ contains mainly lath martensite, with some ferrite and bainite. The amount of ferrite and bainite increases as the heat input increases from 32.5 to 92.22 J/mm, reducing microhardness from 399 to 358 HV0.1. While the softened HAZ microstructure contains tempered martensite and ferrite, the increase of heat input causes more severe martensite tempering with carbide precipitation within grains and grain boundaries.

  4. 4.

    All WJs have joint efficiencies ranging from 92 to 110%, indicating higher WJ quality, while the maximum uniform elongation was found only 57% of the BM. Due to HAZ softening, the yield, tensile strength and uniform elongation of WJs decreased with increase of heat input.

  5. 5.

    During tensile deformation, the formation of void sheet occurred in both high and low heat input WJs. Tempered martensite cracking and void initiation at heavily strained tempered martensite were the predominant mechanisms for void nucleation in high heat input weld specimen. In contrast, for low heat input weld specimen, spherical void nucleation at the interface of ferrite and tempered martensite was observed to be the active damage initiation mechanism during tensile deformation.