Microstructure and Mechanical Properties of Laser-Welded DP Steels Used in the Automotive Industry

The aim of this work was to investigate the microstructure and the mechanical properties of laser-welded joints combined of Dual Phase DP800 and DP1000 high strength thin steel sheets. Microstructural and hardness measurements as well as tensile and fatigue tests have been carried out. The welded joints (WJ) comprised of similar/dissimilar steels with similar/dissimilar thickness were consisted of different zones and exhibited similar microstructural characteristics. The trend of microhardness for all WJs was consistent, characterized by the highest value at hardening zone (HZ) and lowest at softening zone (SZ). The degree of softening was 20 and 8% for the DP1000 and DP800 WJ, respectively, and the size of SZ was wider in the WJ combinations of DP1000 than DP800. The tensile test fractures were located at the base material (BM) for all DP800 weldments, while the fractures occurred at the fusion zone (FZ) for the weldments with DP1000 and those with dissimilar sheet thicknesses. The DP800-DP1000 weldment presented similar yield strength (YS, 747 MPa) and ultimate tensile strength (UTS, 858 MPa) values but lower elongation (EI, 5.1%) in comparison with the DP800-DP800 weldment (YS 701 MPa, UTS 868 MPa, EI 7.9%), which showed similar strength properties as the BM of DP800. However, the EI of DP1000-DP1000 weldment was 1.9%, much lower in comparison with the BM of DP1000. The DP800-DP1000 weldment with dissimilar thicknesses showed the highest YS (955 MPa) and UTS (1075 MPa) values compared with the other weldments, but with the lowest EI (1.2%). The fatigue fractures occurred at the WJ for all types of weldments. The DP800-DP800 weldment had the highest fatigue limit (348 MPa) and DP800-DP1000 with dissimilar thicknesses had the lowest fatigue limit (<200 MPa). The fatigue crack initiated from the weld surface.


Introduction
Advanced high strength steel (AHSS) is the fastest growing material group in today's automotive industry, because of its high strength to weight ratio performance, which allows the car makers to produce thinner components and, thereby, reduce the fuel consumption [1][2][3]. Dual phase (DP) steel group is among the AHSS family widely used in the crash zones of the vehicle due to its high energy absorption ability [3][4][5]. During the past two decades, as the most essential joining technique used in the automotive industry, laser welding has become popular because of its ability to increase the production rate and allow great flexibility of the joint design, without increasing the metallurgical heterogeneities across the weldments [6,7].
The interest has increased on the welding of dissimilar materials with the purpose to reduce both production and operation costs and optimize the properties, such as strength and hardness. Considering the advantages of the laser welding, such as the small heat-affected zone, low heat input, small distortion level etc. [6,7], it is attracting more attentions due to its potential to achieve this purpose. However, during the welding process, element mixing  Three types of tubes were tested, differing in type and wall thickness. The buttwelded joints were produced using a single-pass single-beam laser weld using an IPG fiber laser (YRL-15000, cw, IPG Laser GmbH, Carl-Benz-Straße 28, 57299 Burbach, Germany) in combination with a Precitec YW52 optics (by Precitec GmbH & Co. KG, Draisstraße 1, 76571 Gaggenau, Germany), with the welding parameters presented in Table 3. The line energy is 60 J/mm. The setup of the laser welding experiments used are shown in Figure 1. Laser welding was performed in the focal position with a laser beam angel of 7 • applying argon (purity 99.998%) as shielding gas at 18 L/min to cover the processing zone from the surrounding air. The square tubes to be welded were clamped in a "zero-" gap configuration. No additional edge preparation was used to enhance fitting. One weld seam was produced for each parameter set. The material was cleaned using ethanol before welding to avoid impacts of contamination.   A typical WJ on a tube is shown in Figure 2a. For mechanical testing, the sides 1, 3 and 4 were used due to the tube weld seam alongside 2 ( Figure 2b). For minimization of distortion effects on subsequent welding, the welding sequence in Figure 2b was chosen. Visual testing was conducted in order to identify visible defects such as surface cracks or lack of fusion. However, the produced weld seams showed sound appearances.

Microstructure Investigation and Mechanical Testing
The specimens used in the microstructure and microhardness investigations were grounded, polished and etched in 3% Nital solution to reveal the microstructures. Microstructure investigation was performed using optical microscopy (OM) Nikon Eclipse MA200 (Nikon Instech Co., Ltd., Tokyo, Japan, Japan). Vickers microhardness test was performed at load of a 500 g and a dwell time of 15 s across the base metal, HAZ, and weld metal. The fatigue fracture surfaces were examined with a scanning electron microscopy (SEM) JEOL JSM-IT300LV (JEOL Ltd. Tokyo, Japan). A typical WJ on a tube is shown in Figure 2a. For mechanical testing, the sides 1, 3 and 4 were used due to the tube weld seam alongside 2 ( Figure 2b). For minimization of distortion effects on subsequent welding, the welding sequence in Figure 2b was chosen.  A typical WJ on a tube is shown in Figure 2a. For mechanical testing, the sides 1, 3 and 4 were used due to the tube weld seam alongside 2 ( Figure 2b). For minimization of distortion effects on subsequent welding, the welding sequence in Figure 2b was chosen. Visual testing was conducted in order to identify visible defects such as surface cracks or lack of fusion. However, the produced weld seams showed sound appearances.

Microstructure Investigation and Mechanical Testing
The specimens used in the microstructure and microhardness investigations were grounded, polished and etched in 3% Nital solution to reveal the microstructures. Microstructure investigation was performed using optical microscopy (OM) Nikon Eclipse MA200 (Nikon Instech Co., Ltd., Tokyo, Japan, Japan). Vickers microhardness test was performed at load of a 500 g and a dwell time of 15 s across the base metal, HAZ, and weld metal. The fatigue fracture surfaces were examined with a scanning electron microscopy (SEM) JEOL JSM-IT300LV (JEOL Ltd. Tokyo, Japan). Visual testing was conducted in order to identify visible defects such as surface cracks or lack of fusion. However, the produced weld seams showed sound appearances.

Microstructure Investigation and Mechanical Testing
The specimens used in the microstructure and microhardness investigations were grounded, polished and etched in 3% Nital solution to reveal the microstructures. Microstructure investigation was performed using optical microscopy (OM) Nikon Eclipse MA200 (Nikon Instech Co., Ltd., Tokyo, Japan, Japan). Vickers microhardness test was performed at load of a 500 g and a dwell time of 15 s across the base metal, HAZ, and weld metal. The fatigue fracture surfaces were examined with a scanning electron microscopy (SEM) JEOL JSM-IT300LV (JEOL Ltd. Tokyo, Japan).
Tensile and fatigue test samples were prepared following the ASTM E466-96 standard [27] with the geometry and dimensions shown in Figure 3. The samples were sectioned from the welded work pieces with the weld positioned at the center of the gauge length. The fatigue tests were carried out using Instron servo-hydraulic universal testing machine (Instron. Co. UK., High Wycombe, UK) equipped with data acquisition and hydraulic pressure grips at a frequency of 20 Hz and with a stress ratio of 0.1.
Tensile and fatigue test samples were prepared following the ASTM E466-96 standard [27] with the geometry and dimensions shown in Figure 3. The samples were sectioned from the welded work pieces with the weld positioned at the center of the gauge length. The fatigue tests were carried out using Instron servo-hydraulic universal testing machine (Instron. Co. UK., High Wycombe, UK) equipped with data acquisition and hydraulic pressure grips at a frequency of 20 Hz and with a stress ratio of 0.1.

Appearance of Welded Joints and Microhardness Results
The cross-sections of the different types of WJs are displayed in Figure 4. The results showed that, under the same laser welding conditions, the weld center of DP1000/1.3-DP1000/1.3 and DP800/1.3-DP1000/1.3 were convex, while concave weld center was observed in the DP800/2.1-DP800/2.1 and DP800/2.1-DP1000/1.3. These weld centers were different from that of DP800/1.3-DP800/1.3, for which a flat weld center was observed. The different weld seam appearances may result from slight alignment differences due to the clamping in technical and not nominal zero-gap configuration. A slight gap can lead to a concave seam appearance.

Appearance of Welded Joints and Microhardness Results
The cross-sections of the different types of WJs are displayed in Figure 4. The results showed that, under the same laser welding conditions, the weld center of DP1000/1.3-DP1000/1.3 and DP800/1.3-DP1000/1.3 were convex, while concave weld center was observed in the DP800/2.1-DP800/2.1 and DP800/2.1-DP1000/1.3. These weld centers were different from that of DP800/1.3-DP800/1.3, for which a flat weld center was observed. The different weld seam appearances may result from slight alignment differences due to the clamping in technical and not nominal zero-gap configuration. A slight gap can lead to a concave seam appearance.
Tensile and fatigue test samples were prepared following the ASTM E466-96 stand-ard [27] with the geometry and dimensions shown in Figure 3. The samples were sectioned from the welded work pieces with the weld positioned at the center of the gauge length. The fatigue tests were carried out using Instron servo-hydraulic universal testing machine (Instron. Co. UK., High Wycombe, UK) equipped with data acquisition and hydraulic pressure grips at a frequency of 20 Hz and with a stress ratio of 0.1.

Appearance of Welded Joints and Microhardness Results
The cross-sections of the different types of WJs are displayed in Figure 4. The results showed that, under the same laser welding conditions, the weld center of DP1000/1.3-DP1000/1.3 and DP800/1.3-DP1000/1.3 were convex, while concave weld center was observed in the DP800/2.1-DP800/2.1 and DP800/2.1-DP1000/1.3. These weld centers were different from that of DP800/1.3-DP800/1.3, for which a flat weld center was observed. The different weld seam appearances may result from slight alignment differences due to the clamping in technical and not nominal zero-gap configuration. A slight gap can lead to a concave seam appearance.  A typical DP800/1.3-DP800/1.3WJ was selected to demonstrate the different zones in the WJ, since similar characteristics were observed among the different types of WJs, namely fusion zone (FZ), hardening zone (HZ) and softening zone (SZ), as illustrated in Figure 5a. The microhardness along the different zones were measured, as shown in Figure  5b. The hardness in the FZ displayed an approximate value of 370 HV, while the highest hardness appeared at the boundary between the FZ and HZ. A hardness drop was found in the HZ, the minimum value around 220 HV appeared in the SZ. A hardness increase was observed from the SZ to the BM region.
A typical DP800/1.3-DP800/1.3WJ was selected to demonstrate the different zones in the WJ, since similar characteristics were observed among the different types of WJs, namely fusion zone (FZ), hardening zone (HZ) and softening zone (SZ), as illustrated in Figure 5a. The microhardness along the different zones were measured, as shown in Figure 5b. The hardness in the FZ displayed an approximate value of 370 HV, while the highest hardness appeared at the boundary between the FZ and HZ. A hardness drop was found in the HZ, the minimum value around 220 HV appeared in the SZ. A hardness increase was observed from the SZ to the BM region. The microhardness profiles across the five types of WJs are shown in Figure 6. Consistent trends were found for the different types of WJs, featuring peaks in the HZ and valleys in the SZ. The lowest microhardness values in DP800/1.3-DP800/1.3 and DP1000/1.3-DP1000/1.3 WJs were between 220 and 275 HV, which was 8 and 20% lower in comparison with the corresponding base material. In addition, the width of SZ in DP800/1.3-DP800/1.3 WJ was around 0.49 mm, narrower compared with the DP1000/1.3-DP1000/1.3 WJ with the value of ~0.75 mm. This agrees with the results reported that the softening degree and the size of soften zone was increased with increasing strength grade of the steels [17,18]. The DP800/2.1-DP800/2.1 WJ had wider SZ (~0.54 mm) and higher microhardness in the FZ compared with DP800/1.3-DP800/1.3 WJ. A comparison of the WJs for combinations with dissimilar materials with those of similar materials showed that the SZ of DP800/1.3-DP1000/1.3 WJ at DP800 side was wider, while similar width was observed at the DP1000 side.  The microhardness profiles across the five types of WJs are shown in Figure 6. Consistent trends were found for the different types of WJs, featuring peaks in the HZ and valleys in the SZ. The lowest microhardness values in DP800/1.3-DP800/1.3 and DP1000/1.3-DP1000/1.3 WJs were between 220 and 275 HV, which was 8 and 20% lower in comparison with the corresponding base material. In addition, the width of SZ in DP800/1.3-DP800/1.3 WJ was around 0.49 mm, narrower compared with the DP1000/1.3-DP1000/1.3 WJ with the value of~0.75 mm. This agrees with the results reported that the softening degree and the size of soften zone was increased with increasing strength grade of the steels [17,18]. The DP800/2.1-DP800/2.1 WJ had wider SZ (~0.54 mm) and higher microhardness in the FZ compared with DP800/1.3-DP800/1.3 WJ. A comparison of the WJs for combinations with dissimilar materials with those of similar materials showed that the SZ of DP800/1.3-DP1000/1.3 WJ at DP800 side was wider, while similar width was observed at the DP1000 side.
A typical DP800/1.3-DP800/1.3WJ was selected to demonstrate the different zones in the WJ, since similar characteristics were observed among the different types of WJs, namely fusion zone (FZ), hardening zone (HZ) and softening zone (SZ), as illustrated in Figure 5a. The microhardness along the different zones were measured, as shown in Figure 5b. The hardness in the FZ displayed an approximate value of 370 HV, while the highest hardness appeared at the boundary between the FZ and HZ. A hardness drop was found in the HZ, the minimum value around 220 HV appeared in the SZ. A hardness increase was observed from the SZ to the BM region. The microhardness profiles across the five types of WJs are shown in Figure 6. Consistent trends were found for the different types of WJs, featuring peaks in the HZ and valleys in the SZ. The lowest microhardness values in DP800/1.3-DP800/1.3 and DP1000/1.3-DP1000/1.3 WJs were between 220 and 275 HV, which was 8 and 20% lower in comparison with the corresponding base material. In addition, the width of SZ in DP800/1.3-DP800/1.3 WJ was around 0.49 mm, narrower compared with the DP1000/1.3-DP1000/1.3 WJ with the value of ~0.75 mm. This agrees with the results reported that the softening degree and the size of soften zone was increased with increasing strength grade of the steels [17,18]. The DP800/2.1-DP800/2.1 WJ had wider SZ (~0.54 mm) and higher microhardness in the FZ compared with DP800/1.3-DP800/1.3 WJ. A comparison of the WJs for combinations with dissimilar materials with those of similar materials showed that the SZ of DP800/1.3-DP1000/1.3 WJ at DP800 side was wider, while similar width was observed at the DP1000 side.

Microstructures
Figures 7 and 8 present optical micrographs taken at typical locations in the WJs. As shown in the figures, different zones in the weld metals can be distinguished. For each specific zone, all WJs exhibited similar microstructural constituents and morphologies. Therefore, a representative DP800/1.3-DP800/1.3 WJ and a DP1000/1.3-DP1000/1.3 WJ were selected to illustrate the microstructures of the different zones in the WJs.  The BM DP800 was characterized by martensitic islands in a ferrite matrix, as shown in Figure 7b. The microstructures at the SZ were composed of coarse grains of ferrite and probably tempered martensite as shown in Figure 7c, with minimum hardness values in the hardness profiles. The volume fraction of ferrite towards the FZ was decreased, indicating partial transformation of ferrite. Martensite and ferrite at the HZ were observed, and the grain size was bigger near the FZ (Figure 7d). The FZ was dominated by lath martensite, and the grain size was larger compared to that in BM (Figure 7e). Similar to DP800/1.3-DP800/1.3 WJ, different zones can be distinguished in DP1000/1.3-DP1000/1.3 WJ and similar microstructure characteristics at specific zones were found, as shown in Figure 8. The observed volume fraction of martensite was higher in the BM of DP1000 compared with DP800 (Figure 8b). A lower volume fraction of ferrite was also observed in the SZ in comparison with BM (Figure 8c). The HZ was dominated by martensite with varying grain size (Figure 8d) and lath martensite was found in the FZ (Figure 8e).

Microstructures
Figures 7 and 8 present optical micrographs taken at typical locations in the WJs. As shown in the figures, different zones in the weld metals can be distinguished. For each specific zone, all WJs exhibited similar microstructural constituents and morphologies. Therefore, a representative DP800/1.3-DP800/1.3 WJ and a DP1000/1.3-DP1000/1.3 WJ were selected to illustrate the microstructures of the different zones in the WJs.  The BM DP800 was characterized by martensitic islands in a ferrite matrix, as shown in Figure 7b. The microstructures at the SZ were composed of coarse grains of ferrite and probably tempered martensite as shown in Figure 7c, with minimum hardness values in the hardness profiles. The volume fraction of ferrite towards the FZ was decreased, indicating partial transformation of ferrite. Martensite and ferrite at the HZ were observed, and the grain size was bigger near the FZ (Figure 7d). The FZ was dominated by lath martensite, and the grain size was larger compared to that in BM (Figure 7e). Similar to DP800/1.3-DP800/1.3 WJ, different zones can be distinguished in DP1000/1.3-DP1000/1.3 WJ and similar microstructure characteristics at specific zones were found, as shown in Figure 8. The observed volume fraction of martensite was higher in the BM of DP1000 compared with DP800 (Figure 8b). A lower volume fraction of ferrite was also observed in the SZ in comparison with BM (Figure 8c). The HZ was dominated by martensite with varying grain size ( Figure 8d) and lath martensite was found in the FZ (Figure 8e). The BM DP800 was characterized by martensitic islands in a ferrite matrix, as shown in Figure 7b. The microstructures at the SZ were composed of coarse grains of ferrite and probably tempered martensite as shown in Figure 7c, with minimum hardness values in the hardness profiles. The volume fraction of ferrite towards the FZ was decreased, indicating partial transformation of ferrite. Martensite and ferrite at the HZ were observed, and the grain size was bigger near the FZ (Figure 7d). The FZ was dominated by lath martensite, and the grain size was larger compared to that in BM (Figure 7e). Similar to DP800/1.3-DP800/1.3 WJ, different zones can be distinguished in DP1000/1.3-DP1000/1.3 WJ and similar microstructure characteristics at specific zones were found, as shown in Figure 8. The observed volume fraction of martensite was higher in the BM of DP1000 compared with DP800 (Figure 8b). A lower volume fraction of ferrite was also observed in the SZ in comparison with BM (Figure 8c). The HZ was dominated by martensite with varying grain size ( Figure 8d) and lath martensite was found in the FZ (Figure 8e).

Tensile Test Results
The representative engineering stress versus engineering strain plots of the tensile tested weldments are displayed in Figure 9. The stress-strain curve of the DP800/1.3-DP1000/1.3 weldment was located between that of DP800/1.3-DP800/1.3 and DP1000/1.3-DP1000/1.3 weldments.

Fatigue Test Results
Fatigue test results of the weldments obtained at R = 0.1, 20 Hz, and room temperature (RT) are plotted in Figure 10. The limit for run out (RO) was set to 1 million cycles (10 6 ). The stress level target for the RO was set to 200 MPa. The results show that, although the points were scattered, probably due to the defects introduced during the welding process, it still can be observed that the DP800/2.1-DP800/2.1 weldment had a lower fatigue life than the DP800/1.3-DP800/1.3 weldment at both high and low levels of stress. An interesting finding in the comparison of these two weldments is that the thicker sheet combination (2.1 mm) shows lower scatter of the results in comparison with the thinner sheet (1.3 mm). The DP800/2.1-DP1000/1.3 weldments exhibited the lowest fatigue life in comparison with the other weldments with a RO stress below 200 MPa. tic deformation.

Fatigue Test Results
Fatigue test results of the weldments obtained at R = 0.1, 20 Hz, and room temperature (RT) are plotted in Figure 10. The limit for run out (RO) was set to 1 million cycles (10 6 ). The stress level target for the RO was set to 200 MPa. The results show that, although the points were scattered, probably due to the defects introduced during the welding process, it still can be observed that the DP800/2.1-DP800/2.1 weldment had a lower fatigue life than the DP800/1.3-DP800/1.3 weldment at both high and low levels of stress. An interesting finding in the comparison of these two weldments is that the thicker sheet combination (2.1 mm) shows lower scatter of the results in comparison with the thinner sheet (1.3 mm). The DP800/2.1-DP1000/1.3 weldments exhibited the lowest fatigue life in comparison with the other weldments with a RO stress below 200 MPa. The fatigue limits (ROs) are tabulated in Table 4. It is shown that the fatigue limit of the DP800/1.3-DP800/1.3 weldment was 348 MPa, which was significantly higher than that of the DP1000/1.3-DP1000/1.3 and DP800/1.3-DP1000/1.3 weldments with the value of approximate 210 MPa. Even though the same material was used, the fatigue limit of DP800/2.1-DP800/2.1 weldment was obtained to be 221 MPa, lower than DP800/1.3-DP800/1.3 weldment.
The typical fatigue failure locations are shown in Figure 11. It was observed that the fracture occurred at the WJ. All types of weldments presented the same behavior. The fatigue limits (ROs) are tabulated in Table 4. It is shown that the fatigue limit of the DP800/1.3-DP800/1.3 weldment was 348 MPa, which was significantly higher than that of the DP1000/1.3-DP1000/1.3 and DP800/1.3-DP1000/1.3 weldments with the value of approximate 210 MPa. Even though the same material was used, the fatigue limit of DP800/2.1-DP800/2.1 weldment was obtained to be 221 MPa, lower than DP800/1.3-DP800/1.3 weldment.
The typical fatigue failure locations are shown in Figure 11. It was observed that the fracture occurred at the WJ. All types of weldments presented the same behavior.

Fractography
SEM images of fatigue fracture surface of a DP800/1.3-DP800/1.3 weldment tested under a maximum stress of 626 MPa are shown in Figure 12. Multiple crack initiation points were found, located at the surface, as shown in Figure 12b. Figure 12c illustrates the fatigue striations in the crack propagation region. The dimples in the final fast crack propagation region indicated the plastic fracture (Figure 12d).

Fractography
SEM images of fatigue fracture surface of a DP800/1.3-DP800/1.3 weldment tested under a maximum stress of 626 MPa are shown in Figure 12. Multiple crack initiation points were found, located at the surface, as shown in Figure 12b. Figure 12c illustrates the fatigue striations in the crack propagation region. The dimples in the final fast crack propagation region indicated the plastic fracture (Figure 12d).

Fractography
SEM images of fatigue fracture surface of a DP800/1.3-DP800/1.3 weldment tested under a maximum stress of 626 MPa are shown in Figure 12. Multiple crack initiation points were found, located at the surface, as shown in Figure 12b. Figure 12c illustrates the fatigue striations in the crack propagation region. The dimples in the final fast crack propagation region indicated the plastic fracture (Figure 12d).   Figure 13. Crack initiation from the surface of the weldment was observed. Impurity particles that appeared on the surface were also noticed. Fatigue striations were found in the crack propagation region, as shown Figure 13c. The obvious dimples at the fracture surface indicated the plastic fractures, meanwhile brittle fractures were noticed (Figure 13d). SEM images of a fatigue fracture surface of a DP1000/1.3-DP1000/1.3 weldment tested under the maximum stress level of 282 MPa are shown in Figure 13. Crack initiation from the surface of the weldment was observed. Impurity particles that appeared on the surface were also noticed. Fatigue striations were found in the crack propagation region, as shown Figure 13c. The obvious dimples at the fracture surface indicated the plastic fractures, meanwhile brittle fractures were noticed (Figure 13d).

Microstructure Evolution
Different microstructural characteristics observed in the WJ could be a result from the different thermal cycles experienced during the welding processes due to the different absorption behavior of the different steel alloys and the differences in heat transport of the varying material geometries. The formation of large martensitic laths at the FZ is attributed to the large amount of heat and subsequent rapid cooling [13,18]. The peak temperature gradually decreases towards the BM, dividing the HAZ into HZ and SZ. As the HZ experienced thermal cycles with a peak temperature higher than Ac3, the ferrite and martensite in the BM were completely converted into austenite during the welding, and martensite is formed during the rapid cooling process. Furthermore, the decreasing peak temperature towards BM leads to the decreasing grain size in the HZ. The peak temperature decreases below the Ac3 with increasing distance from the weld center, leading to the formation of SZ. At the SZ close to the side of HZ, a fraction of ferrite and martensite is transformed into austenite, and the austenite part is converted into martensite and ferrite during the slow cooling cycle. As the temperature decreases further away from the FZ, the temperature will not reach Ac1, and tempered martensite is formed in the ferritic-martensitic structure.

Microstructure Evolution
Different microstructural characteristics observed in the WJ could be a result from the different thermal cycles experienced during the welding processes due to the different absorption behavior of the different steel alloys and the differences in heat transport of the varying material geometries. The formation of large martensitic laths at the FZ is attributed to the large amount of heat and subsequent rapid cooling [13,18]. The peak temperature gradually decreases towards the BM, dividing the HAZ into HZ and SZ. As the HZ experienced thermal cycles with a peak temperature higher than Ac 3 , the ferrite and martensite in the BM were completely converted into austenite during the welding, and martensite is formed during the rapid cooling process. Furthermore, the decreasing peak temperature towards BM leads to the decreasing grain size in the HZ. The peak temperature decreases below the Ac 3 with increasing distance from the weld center, leading to the formation of SZ. At the SZ close to the side of HZ, a fraction of ferrite and martensite is transformed into austenite, and the austenite part is converted into martensite and ferrite during the slow cooling cycle. As the temperature decreases further away from the FZ, the temperature will not reach Ac 1 , and tempered martensite is formed in the ferritic-martensitic structure.

Mechanical Properties
The fracture location and tensile test properties of the BM for DP800/1.3-DP800/1.3 and DP800/2.1-DP800/2.1 weldments indicate no significant effect of the WJ, which is ascribed to the narrow SZ and low degree of softening. A narrow SZ can be better restrained by the neighboring structure [28]. The optimized welding parameters, i.e., faster welding speed and lower laser power, could lead to a narrower softening zone and com-press the effect of SZ on the strength [15,20]. Differently from DP800/1.3-DP800/1.3 and DP800/2.1-DP800/2.1 weldments, the yielding firstly occurred at the SZ for the DP1000/1.3-DP1000/1.3 weldment, due to the much lower microhardness values in the SZ when compared with the BM. The plastic deformation subsequently concentrates there, as failure was observed at the WJ. This, in turn, leads to reduction of the elongation. Therefore, the fracture occurring at the WJ and the tensile property in the DP800/2.1-DP1000/1.3 could be caused by the narrower soft zone and smaller extent of the SZ at the DP800 side while wider extention of the SZ was observed at DP1000 side. This could also explain the higher elongation and fracture location in the DP800/1.3-DP1000/1.3 weldment where the wider SZ at the DP1000 side could accommodate larger strain before failure compared to the narrower SZ in the DP800/2.1-DP1000/1.3 weldment.
The different locations of fracture after tensile and fatigue tests indicates that the narrow SZ does not affect the tensile properties, while the fatigue resistance was sensitive to the presence of the SZ. Although higher fatigue strength was reported in the DP steel with increasing martensite content [29][30][31], lower fatigue strength of DP1000/1.3-DP1000/1.3 weldment than DP800/1.3-DP800/1.3 weldment was found in this work, which could be a result of different surface defects as oxide particles appearing on the weld surface. In addition, the weld concavity can initiate formation of fatigue cracks [32,33]. The lower fatigue strength of DP800/2.1-DP800/2.1 weldment compared with DP800/1.3-DP800/1.3 weldment can be related to the presence of the weld concavity. All these findings confirm that not only the occurrence of SZ, but especially the defects introduced during welding process plays an important role on the fatigue fracture.

Conclusions
Microstructure of the welded joints and the effect of the welding on the tensile and fatigue properties for different combinations of similar/dissimilar materials with different thicknesses were investigated. Based on the results achieved, the following conclusions can be drawn: The evolution trends of microhardness of the five types of the welded joints were consistent, featuring highest values in the fusion zone and lowest values in the soft zone. However, the degree of softening was more severe and the size of the SZ was wider for DP1000 steel than for the DP800 steel.
(1) A difference in thickness, (1.3 to 2.1 mm), for the welded DP800 steel, did not affect the tensile property significantly and the fracture occurred at the base material.