ISIJ International
Online ISSN : 1347-5460
Print ISSN : 0915-1559
ISSN-L : 0915-1559
Welding and Joining
Fusion Zone Microstructural Evolution of Al-10% Si Coated Hot Stamping Steel during Laser Welding
Chang Wook Lee So Youn KimSoon Geun JangGa Young ShinJi Hong Yoo
Author information
JOURNAL OPEN ACCESS FULL-TEXT HTML

2019 Volume 59 Issue 1 Pages 136-143

Details
Abstract

Fusion zone microstructural evolution of Al-10% Si coated hot stamping steel during laser welding and its mechanical properties were investigated in this study. During laser welding, a liquidized Al-10% Si coating penetrated along the fusion boundary to form δ-ferrite. After hot stamping heat treatment, the final microstructure was composed of a martensite-ferrite dual phase. The dual phase formation causes mixed mode fracture (brittle fracture + ductile fracture) at the fusion zone due to the hardness difference between the martensite and ferrite phases. This issue can be addressed by applying filler wire or by changing the coating system from an Al-based coating to a Zn-based coating.

1. Introduction

Use of high strength steel in car bodies has increased in an attempt to achieve improved passenger safety and fuel efficiency. The most efficient way to produce ultra-high strength steel (UHSS) with excellent formability is using a hot stamping process.1) During hot stamping, steel blanks are heated and kept above the Ac3 austenitization temperature. After heated steel blanks are transferred to the forming press, press forming and quenching are simultaneously carried out by water-cooled dies to obtain a full martensite microstructure, generally resulting in a yield strength (YS) of 1.0 GPa and ultimate tensile strength (UTS) of 1.5 GPa.2) Hot stamped steel parts were first used in automotive applications in 1984 when hot-stamped door intrusion beams were used in SAAB 9000 passenger cars.2,3) In 2002, the first pre-coated hot-stamped steel was used in Citroën C5.5) In current manufacturing processes, tailored properties of hot stamped automotive parts are obtained using tailor welded blanks (TWB), tailor rolled blanks (TRB), partial heating, differential cooling and other methods. Because of the continuous and innovative development of hot stamping technology, the number of hot -stamped components has increased considerably, reaching 250 million parts in 2015.6)

When hot stamping technology was first implemented, bare steel was used. During the hot stamping heat treatment, iron oxide scale formed and decarburization occurred at the surface of steel sheets.7,8) Various coatings have therefore been developed to suppress both oxidation and decarburization. The most commonly used coating is the Al-10% Si alloy coating, which provides excellent barrier protection. However, the cathodic protection of this alloy coating is limited.9,10) When the steel substrate is exposed to air by chipping or flaking, the coating cannot protect the steel substrate. A Zn-based coating has been developed to provide cathodic protection for hot-stamped steel.11) The final microstructure of the Zn coating after a hot stamping process is composed of a Γ-Fe3Zn10 intermetallic compound (IMC) and a solid solution of Zn in α-Fe(Zn). The Γ-Fe3Zn10 is beneficial for cathodic protection, but can also cause liquid metal embrittlement (LME). The two main requisites for Zn-LME are the presence of liquid Zn and an applied stress. To suppress LME, a diffusion annealing treatment to eliminate Fe3Zn10 IMC or an indirect hot stamping process can be used.12,13)

Application of TWB is beneficial in that the mechanical properties can be locally adjusted by a proper choice of the joining partner.14) A blank can consist of different strength grades and sheet thicknesses, a feature that cannot be used with other methods to tailor properties. Furthermore, the fusion zone and heat-affected zone (HAZ) are composed of a full martensite microstructure due to additional heat treatment conducted above the A3 temperature during hot stamping heat treatment.15,16,17) In a TWB process, two different steel blanks are laser-welded together. The Al-10% Si coating of the hot stamping steel melted and alloyed with the molten weld pool during laser welding. Since Al is a strong ferrite stabilizer, its incorporation into the fusion zone of laser-welded Al-10% Si-coated blanks following a hot stamping process can result in a large grain size of ferrite phase surrounded by martensite microstructure.15) In this contribution, the effect of an Al-10% Si coating on microstructure evolution during laser welding and a hot stamping process was investigated. The ferrite-martensite dual phase fusion zone caused deterioration of the mechanical properties of hot-stamped TWB parts. In addition, solutions for preventing ferrite-martensite dual phase formation in the fusion zone were evaluated.

2. Experimental

A 22MnB5 (Fe-0.23% C-1.2% Mn-0.004% B in mass%) hot stamping steel grade was prepared by vacuum induction melting (VIM). The ingots were hot rolled prior to cold rolling to final thickness of 1.2 and 1.4 mm. The cold rolled panels were coated separately with two different hot-dip coating methods which are an Al-10% Si alloy coating and a galvannealed (GA) coating. The thickness of the Al-10% Si coating was about 18 μm and that of the GA coating was about 9 μm. The initial microstructures of the Al-10% Si coating and GA coating are shown in Fig. 1. The Al-10% Si coating was composed of an Al–Si alloy layer and a Fe2SiAl7 IMC layer. Most of the GA coating was composed of a δ-FeZn10 IMC phase. SWRH62A (JIS G3506) high C filler was used to convert the make fusion zone into a fully martensite microstructure after hot stamping, as it is composed of 0.66 wt.% C and 0.49 wt.% Mn which are the strong austenite stabilizers.

Fig. 1.

Initial coating conditions of the Al-10%Si coating and galvannealed coating prior to laser welding and hot stamping heat treatment.

A Trumpf TruDisk 8002 solid state laser beam (wavelength of 1030 nm and maximum power of 8.0 kW) and Fronius VR4000 filler supplier were used for laser welding. Two different thickness combinations were used to produce laser-welded blanks: 1.2 mm and 1.2 mm (same thickness), and 1.4 mm and 1.2 mm (different thicknesses). The output power was fixed at 4 kW. The welding speed was 4 m/min when joining sheets of the same thickness and 3 m/min when joining sheets of different thicknesses. The feeding speed for filler wire was 2 m/min, as shown in Table 1.

Table 1. Laser welding conditions according to thickness combination and coating type.
Thickness combinationCoatingFiller wireLaser OutputWelding SpeedFeeding Speed
1.2 t–1.2 tAl-10%SiNo4 kW4 m/min
Al-10%SiYes4 kW4 m/min2 m/min
GANo4 kW4 m/min
1.4 t–1.2 tAl-10%SiNo4 kW3 m/min
GANo4 kW3 m/min

The welded steel blanks (300 mm × 300 mm) were heat treated in a furnace at 950°C for 5 mins with Al-10% Si coated steel and 900°C for 5 mins with galvannealed steel. And then quenched by water-cooled flat dies with an average cooling rate of −50°C/s.

JIS Z2201 No. 5 tensile specimens were prepared. The laser welding line was located at the center of the tensile specimen. The tensile properties were measured at room temperature at a strain rate of 10−3 s−1 using a Zwick Z100 universal tensile testing machine.

Fractography of the tensile testing specimens was carried out. TWB hot stamping steel before and after heat-treatment was cold-mounted in a polymer resin. After polishing with an alcohol-based diamond suspension, the surfaces of the samples were platinum-coated by physical vapor deposition (PVD) to improve their surface conductivity. The cross-sectional fusion zone microstructure was analyzed using a ZEISS Axio Imager. A2m optical microscope (OM) and an FEI Quanta 450 scanning electron microscope (SEM) with energy dispersive spectroscopy (EDS). Cross-sectional TEM samples were extracted from the panel using the focused ion-beam (FIB) milling technique and observed using a FEI Tecnai G2 F30 S-TWIN transmission electron microscope (TEM).

Nanoindentation measurements were conducted using a Hysitron TriboIndenter TI 900 with a Berkovich indenter (Ti 39-01, tip radius 50 nm). Measurements were performed at room temperature in load control mode at a load of 2000 μN. The indentation procedure was programmed as three segments of a trapezoidal shape: 10 s loading, 5 s of holding, and 10 s of unloading.

3. Results & Discussion

Figure 2 shows the uniaxial tensile engineering stress-strain curves for the hot-stamped Al-10% Si-coated 22MnB5 with and without laser welding. The thickness of the specimens was 1.2 mm. Generally, Al-10% Si-coated 22MnB5 hot stamped steel had YS in the range of 1000–1100 MPa, UTS in the range of 1400–1500 MPa and total elongation (TE) in the range of 6.5–7.0%, as shown in Fig. 2. However, laser-welded Al-10% Si-coated 22MnB5 hot-stamped steel had a TE of only 1.4%. The brittle fracture occurred at the welding line during uniform elongation. The laser-welded sample had a 5.4% ductility loss compared with the sample without laser welding.

Fig. 2.

Engineering stress-strain curves of laser-welded and hot-stamped Al-10% Si-coated 22MnB5 steel.

Figure 3 shows the fracture surface of laser-welded Al-10% Si-coated 22MnB5 hot-stamped steel after tensile testing. The fracture mode consisted of cleavage fracture and ductile fracture. In area A, many substantially planar crystallographic cleavage facets were observed. By tracing back the river pattern strips on the fracture surface to their origins, it can be clearly observed that all of the cleavage fractures initiate in the matrix itself without inclusion at the initiation site. In area B, a combination of the cleavage fracture and ductile fracture was observed. The cleavage facets contained 2.9 wt.% Al, which is a strong ferrite stabilizer whereas the ductile area contained 0.7% wt.% Al. The fusion zone was composed of a dual phase microstructure, not a full martensite microstructure due to the high Al content.

Fig. 3.

(a, b) Fracture surface micrograph of laser-welded and hot-stamped Al-10% Si-coated 22MnB5 steel following tensile testing.

Figure 4 shows a cross-sectional micrograph of laser-welded Al-10% Si-coated 22MnB5 steel before and after hot stamping heat treatment. Before heat treatment, the Al-10% Si coating penetrated along the boundary between the fusion zone and HAZ. After heat treatment, Al containing phases were universally distributed in the fusion zone.

Fig. 4.

SEM micrograph of fusion zone of laser-welded Al-10% Si-coated 22MnB5 steel before and after hot stamping heat treatment.

Figure 5 shows a Fe (C, Mn, Si, Cr)-x% Al binary phase diagram created using Thermo-Calc™ software and TCFE7 database to determine the phase transformation with respect to the Al content and base metal elements (0.23 wt.% C, 1.2 wt.% Mn, 0.26 wt.% Si, 0.19% Cr). This figure shows that when the Al content is higher than 2.7 wt.%, a δ-ferrite was a primary forming first from liquid (Liquid → Liquid + Ferrite) and afterward an austenite-ferrite dual phase was formed (Liquid + Ferrite → Austenite + Ferrite) during cooling. The austenite transformed into α’-martensite instead of cementite due to rapid cooling. Therefore, final microstructure was composed of δ-ferrite and α’-martensite after laser welding. In addition, if the Al content is higher than 1.25 wt.% at a hot stamping heat treatment temperature of 950°C, the resulting steel matrix would be composed of an α-ferrite and γ-austenite dual phase via recrystallization, rather than an austenite single phase. After quenching, the final microstructure would comprise α-ferrite and α’-martensite.

Fig. 5.

Fe–Al binary phase diagram for Fe (C, Mn, Si, Cr)-x%Al of 22MnB5 steel grade used in this study, calculated using Thermo-Calc™.

TEM analysis was carried out to confirm that the Al containing phase is ferrite or Fe–Al intermetallic compounds. Figures 6 and 7 show dark-field STEM micrographs, the corresponding selected area diffraction patterns (SADP) and the TEM-EDS compositional analysis. The TEM samples were extracted from at the boundary between martensite and the Al containing phase by FIB milling. Before hot stamping, a mixture of lath martensite and polygonal δ-ferrite microstructure was observed. The body centered tetragonal (BCT) α’-martensite and the body centered cubic (BCC) δ-ferrite phases cannot be distinguished by diffraction patterns. The a lattice parameter was 0.2787 nm and the c lattice parameter was 0.2835 nm. The c lattice parameter is highly dependent on the carbon content. Previous research showed that when the carbon content was 4 wt.%, the difference between the a and c parameters of the BCT structure was about 0.1 nm.18) In this study, the difference between the a and c parameters in the BCT martensite microstructure was 0.0048 nm as the carbon content of the specimen was 0.23 wt.%.

Fig. 6.

STEM DF micrograph and corresponding diffraction patterns, elemental concentration for selected points and elemental distribution maps for Fe, Al, Si, Cr and Mn for the dual phase microstructure in laser-welded Al-10% Si-coated 22MnB5 steel.

Fig. 7.

STEM DF micrograph and corresponding diffraction patterns, elemental concentration for selected points, and elemental distribution maps for Fe, Al, Si, Cr, and Mn for the dual phase microstructure in laser-welded and hot-stamped Al-10% Si-coated 22MnB5 steel.

Before the hot stamping heat treatment, the Al content of δ-ferrite in Fig. 6 was 2.1 wt,% which was relatively lower than the 5.2 wt.% Al content of δ-ferrite shown in Fig. 4 because the specimen was extracted at the tip far from the coating. The δ-ferrite located close to the coating can be provided more Al from the coating. The difference in Al content between δ-ferrite and α’-martensite was only 0.4 wt.%; however, the morphology of the two phases clearly differed. After hot stamping heat treatment, the difference in the Al content between α-ferrite and α’-martensite was 0.7 wt.%. Although the measured Al content difference between ferrite and martensite was small (as the martensite location was close to the Al containing ferrite region), a small amount of Al can cause ferrite formation at high temperatures.

The mechanism of fusion zone microstructure evolution is summarized in the schematic shown in Fig. 8. Initially, the base metal was composed of ferrite and pearlite microstructures. After laser welding, most of the fusion zone was composed of α’-martensite. A relatively small amount of δ-ferrite was formed along the fusion boundary due to molten Al-10%Si coating penetration. During the hot stamping heat treatment, the fusion zone was composed of a γ + α dual phase microstructure instead of an austenite single phase microstructure because concentrated Al at the fusion boundary diffused into the fusion zone and increased the A3 temperature above 950°C. In addition, partitioning of aluminium to the ferrite during the heat treatment leads to stabilization of ferrite.16) After die quenching, the γ + α dual phase transformed to an α’ + α dual phase. The microstructure of the fusion zone after laser welding and hot stamping heat treatment could vary depending on factors such as the material conditions (chemical composition of base metal, coating thickness), laser welding conditions (laser output, laser beam size, welding speed, wire feeding speed) and hot stamping conditions (annealing temperature, soaking time, quenching rate).15,16,17,20,21)

Fig. 8.

Schematic of mechanism for fusion zone microstructure evolution during laser welding and hot stamping processes.

A nanoindentation test was carried out to measure the difference in hardness between the α’-martensite phase and α-ferrite phase. The load-displacement curves and their corresponding indent impressions are shown in Fig. 9. The nanohardness measured on the α’-martensite phase was about 7.0 GPa with an indentation depth of ~65 nm and the nanohardness measured on the α-ferrite phase was 3.7–4.3 GPa with an indentation depth of ~105 nm. The large differences in hardness between α’-martensite and α-ferrite phases causes brittle fracture. In addition, the large grain size of α-ferrite, as shown in Fig. 4, also affects brittle fracture.

Fig. 9.

Nanohardness and load-depth curves of α-ferrite and α’-martensite dual phase microstructure in fusion zone of laser-welded and hot-stamped Al-10% Si-coated 22MnB5 steel.

The fracture of laser-welded and hot-stamped Al-10% Si-coated 22MnB5 steel is caused by dual phase formation due to Al content in the fusion zone. Thus, full austenitization in this zone is required to avoid brittle cleavage fracture. Three main solutions have been applied to solve this issue: removing the Al-10% Si coating, inducing strong austenite stabilizers, and changing the coating system from an Al-10% Si coating to a Zn-based coating.

The first method is removing the Al–Si eutectic phase layer of the Al-10%Si coating using a short multi-impulse laser before laser welding, also known as laser ablation. After laser ablation, only the Fe–Al intermetallic phase layer at the interface between the Al-10%Si coating and steel substrate remains. Moon et al. reported that laser ablation must be performed on both sides of the coating layer. If laser ablation is carried out on only one side for cost saving purposes, the tensile properties and hardness suffer.21)

Secondly, if strong austenite stabilizing elements (such as C and Mn) are added to the fusion zone, the microstructure of the fusion zone will be fully composed of an austenite microstructure at high temperature and a martensite microstructure after die-quenching to room temperature. According to thermodynamic calculations (Fig. 10), when the carbon content increases from 0.2 wt.% to 0.6 wt.%, the microstructure of the fusion zone can become fully austenite at 950°C and fully martensite after die quenching to room temperature when the Al content is below 3.7 wt.%. The most efficient way to incorporate a strong austenite stabilizer during laser welding is by using filler wire. In this study, SWRH62A high carbon steel wire was applied. This steel wire contains 0.66 wt.% C and 0.49 wt,% Mn which is sufficient to induce a fully austenite microstructure in the fusion zone at 950°C. Figure 11 shows the uniaxial tensile engineering stress-strain curves for the hot-stamped laser-welded blank with filler wire of Al-10% Si-coated 22MnB5 steel. The tensile strength and total elongation reached 1546 MPa and 5.5% without any brittle failure during tensile testing. The fracture was not located at the fusion zone. The microstructure of the fusion zone was fully martensite and the formation of α-ferrite could not be observed. Tandon et al. also reported that the filler wire is a key enabler for improving robustness in three ways: controlling weld geometry (gap bridging), aluminum dilution, and introducing austenite stabilizer (C, N).22)

Fig. 10.

Effect of carbon on Fe–Al binary phase diagram for Fe (C, Mn, Si, Cr)-x%Al of 22MnB5 steel grade in this study calculated using Thermo-Calc™.

Fig. 11.

Engineering stress-strain curves of laser-welded with filler wire and hot-stamped Al-10% Si-coated 22MnB5 steel.

Finally, if a Zn-based coating is applied instead of an Al-10% Si coating, brittle fracture due to ferrite formation at the fusion zone could be prevented. Figure 12 shows the uniaxial tensile engineering stress-strain curves for the hot-stamped laser-welded blank with GA-coated 22MnB5 steel. For the laser-welded specimen made from blanks of the same thickness (1.2 mm), tensile strength and total elongation reached 1447 MPa and 2.6% and fracture occurred at the fusion line. From the stress-strain curve, sudden fracture during plastic deformation occurred before the UTS was reached; however, a ductile dimple fracture surface was observed (Fig. 13). Before and after hot stamping, the microstructure of the fusion zone did not contain any ferrite phases. It could be that the fusion zone concavity acts as a notch for fracture during plastic deformation. In the case of the laser-welded blank (GA-coated 22MnB5) with different thicknesses (1.4 mm and 1.2 mm), the tensile strength and total elongation reached 1577 MPa and 5.2%, whereas the Al-Si-coated specimen fracture in the same way as seen in Fig. 2. The fracture did not occur along the laser welding line, but rather at the 1.2 mm specimen because following welding of the two blanks of different thicknesses, most of loading during tensile testing would occur in thinner blank. In addition, when the thickness of two blanks was different, the thickness of the fusion zone gradually decreased from the thicker blank to the thinner blank, as shown in Fig. 14. This thickness compensation is highly dependent on the heat input, which is controlled by welding conditions such as laser output power and welding speed.

Fig. 12.

Engineering stress-strain curves of laser-welded wire and hot-stamped GA 22MnB5 steel.

Fig. 13.

Fracture surface SEM micrograph of tensile tested specimen and cross-sectional OM images of fusion zone of GA 22MnB5 steel before and after hot stamping heat treatment.

Fig. 14.

Engineering stress-strain curves of laser-welded wire and hot-stamped GA 22MnB5 steel using welded blanks with different thicknesses (1.2 mm–1.4 mm).

Zinc is also a strong ferrite stabilizer; however, the boiling temperature of Zn is 907°C. Therefore, during laser welding, Zn vaporized completely, while Al liquidized and penetrated the fusion zone. The Zn-coated blank offers cost saving benefits, as there is no need for additional processing steps for laser ablation to remove the coating or incorporation of filler wire to preventing ferrite formation in the fusion zone. Furthermore, the weld seam could be protected against corrosion by the remote cathodic protection of an Fe–Zn alloyed coating.23)

4. Conclusion

Effect of an Al-10% Si coating on the microstructure and mechanical properties of laser-welded and hot-stamped steel was investigated. The filler wire technique and change of the coating system from an Al-10% Si coating to a galvannealed coating was applied to enhance the stability of the fusion zone to prevent ferrite formation.

The following conclusions can be drawn from the present study:

(1) The mechanical properties of laser-welded and hot-stamped Al-10% Si-coated 22MnB5 steel did not achieve the target properties due to ferrite formation in the fusion zone caused by Al diffusion.

(2) After laser welding, the fusion zone was composed of α’-martensite and δ-ferrite, which formed at the fusion boundary because of molten Al-10% Si coating penetration.

(3) After hot stamping heat treatment, the fusion zone was composed of α’-martensite and α-ferrite. α-ferrite formed in the entire fusion zone due to Al diffusion during hot stamping heat treatment. This ferrite formation caused brittle fracture along the laser welding line.

(4) By applying high carbon filler wire during laser welding, brittle fracture can be prevented. C acted as a strong austenite stabilizer to make fusion zone fully austenite at high temperature and then fully martensite after quenching to room temperature during laser welding and hot stamping processes.

(5) When the Zn-based coating was applied instead of the Al-based coating, the fusion zone did not contain any ferrite phase before and after hot stamping, even though Zn is also a ferrite-stabilized element (like Al). Zn could not penetrate into the fusion zone because Zn was totally vaporized due to its low boiling temperature 907°C. The fusion zone can be transformed from fully austenite at high temperature to fully martensite at room temperature during laser welding and hot stamping processes. It can also be an alternative solution for preventing brittle fracture along the laser welding line.

References
 
© 2019 by The Iron and Steel Institute of Japan
feedback
Top