MATERIALS TRANSACTIONS
Online ISSN : 1347-5320
Print ISSN : 1345-9678
ISSN-L : 1345-9678
Materials Processing
Microstructure of Repair Welding Heat-Affected Zone of a Mo-Modified AISI H13 Hot-Work Tool Steel for Die-Casting Die
Yohei SakurabaRyota KawakamiNaofumi Ohtsu
著者情報
ジャーナル フリー HTML

2025 年 66 巻 3 号 p. 358-365

詳細
Abstract

The heat-affected zone (HAZ) formed on welding-repaired die-casting die features a hardness distribution that can potentially initiate thermal shock cracks. This study aimed to analyze the microstructure of a repair welding HAZ formed on Molybdenum-modified AISI H13 steel (Mo-modified H13 steel) used for the die-casting die in industry to identify the characteristics causing this hardness distribution. In a model weld-repaired die, the HAZ from 0–2.8 mm from the boundary of the weld metal was hardened by up to 140 HV compared with that of the original substrate, whereas the region from 2.8–5.0 mm was softened by up to 200 HV. Microstructural and crystallographic analyses reveal that the hardened region had a martensite structure, which corresponded to the origin of the hardening. Contrastingly, the softened region exhibited a typical annealed structure with granular Mo and V carbides. This structure results in the loss of the secondary hardening effect of the Mo-modified H13 steel substrate, which decreases hardness.

Fig. 2 Hardness distribution in the welding specimen.

1. Introduction

Aluminum alloy die-casting dies are generally made of hot-work tool steel owing to its high-temperature strength, toughness, and wear resistance. During the casting process, the die is in contact with molten metal at ∼973 K and then is quenched below 473 K when releasing the casting product. Thus, the die experiences a large thermal stress derived from the temperature difference of ∼500 K. Such large thermal stresses often induce thermal fatigue cracks in hot-work steel dies, which gradually grow as the casting process repeats [13]. Thermal fatigue cracks cause dimensional errors in casting products; hence, they should be repaired to maintain dimensional accuracy [2, 4]. The generic repair process for hot-work tool steel die involves removing the damaged part and then clad welding material to restore the original shape. Therefore, the formation of a welding heat-affected zone (HAZ) within the original die is unavoidable. It is well known that such HAZs show heterogeneous hardness, which can cause aggravate thermal fatigue cracking [5, 6]. Consequently, the lifetime of the repaired die is significantly shortened. Removal of such hardness heterogeneity by further heat treatment can effectively prolong the lifetime [4, 5, 7]. Nevertheless, it is difficult to determine adequate treatment conditions (temperature and duration) that is generally dependent on a technician’s experience.

To systematically determine an adequate post-treatment condition, the origin of the hardness heterogeneity in the welding HAZ formed on an actual repair welded die-casting die should be analyzed at microstructure scale. AISI H13 steel is primarily used for die-casting die. Braga et al. [8] analyzed the structure of the HAZ in an AISI H13 hot-work tool steel substrate that was formed during laser deposition of powder of the same steel. They demonstrated that the increase in hardness in the HAZ was caused by the refinement of the crystal grains formed by the precipitation of granular V, Mo, Cr, and Fe carbides. In addition, Telasang et al. [9] analyzed the HAZ structure formed by the irradiation of AISI H13 steel using a laser beam with different energy densities and revealed that a martensitic structure appeared in a comparatively high-temperature region, whereas the low-temperature region did not have this structure but rather had a similar structure to the original H13 steel. Chiang et al. [10] examined the structural changes in the HAZ formed by laser melting with reference to a theoretically calculated surface temperature distribution. They revealed that the surface temperature was elevated beyond the austenitic transformation temperature, resulting in a martensitic structure after quenching. In addition, structural changes in AISI H13 steel resulting from annealing and additive manufacturing have also been studied [1117].

To elucidate the HAZ formed by die-repair welding, the HAZ formed on an actual repair welded die should be meticulously analyzed; however, studies on this phenomenon are scarce. Furthermore, in recent years, die-casting dies have been fabricated using modified H13 steel with a modified elemental composition rather than original AISI H13 steel. For instance, Mo-enriched H13 steel is used for aluminum die-casting dies owing to its better hardenability and welding properties. Therefore, in this study, a model die made of a Mo-modified H13 hot-work steel was welded under conditions imitating a die-repair process, after which the welding HAZ was analyzed at the microstructure scale. Based on the results, we explored the principle of the formation of hardness distributions in the welding HAZ from die repair.

2. Experimental Procedure

2.1 Specimen

In this study, DH31-S (Daido Steel Corporation, Japan) plates made of Mo-modified H13 hot-work tool steel with dimensions of 80 mm × 70 mm × 20 mm were used as the original substrate. Additionally, a standard AISI H13 steel plate with a similar size was used for comparison. GS-3 steel (Stehle Corporation, Germany) with a diameter of 0.8 mm was used as the filler metal. The nominal chemical compositions of these steels provided by the manufacturers are listed in Table 1. The composition of the modified steel substrate is similar to that of the standard steel; however, it featured high Mo content and reduced C. Such compositional modification enables us to enhance the secondary hardening effect through Mo carbide formation [18, 19], thereby improve the hardenability and toughness [20]. The Mo-modified H13 hot-work steel and the filler steels were selected as the typical steels used in an actual aluminum die-casting die in industry.

Table 1 Material composition of the substrate and filler metal.


The steel plate substrates were machined to form a U-shaped groove with an area of 80 mm × 36 mm and a depth of 3 mm using a vertical milling machine. The steel plates were then hardened at 1303 K for 1 h, followed by double tempering at 843 K and 878 K for 3 h each, according to ISO 4957(1999) [21]. Upon completing the aforementioned processes, the filler metal was welded into the U-groove of the steel plates. Welding was performed using a TPS4000MV MIG/MAG welder (Fronius International GmbH, Austria). Here, electric current and voltage were set to 160 A and 33 V, respectively, and a mixture of 80 vol% Ar and 20 vol% CO2 at a flow rate of 25 L min−1 was used as the shielding gas. The welding parameters were determined through preliminary experiments aimed at minimizing welding defects. The plate was preheated at 623 K for 1 h before welding, and reheated at 648 K for 1 h immediately after the completion of welding, followed by cooling to room temperature. The plate surfaces were machined using a grinding machine to prepare a flat surface without excess weld metal. Figure 1(a) shows the ground plate of the modified steel. A small sample region including the steel plate and the filler metal was cut, as indicated by the dashed lines in the figure, and the surface was chemically etched for 300 s using a 5 vol% nitric acid solution (Fig. 1(b)). The boundary between the weld metal and the HAZ in the original substrate appeared as a color change from dark gray to black, which was defined as the reference point (0.0 mm). Hereafter, the steel plate containing the filler metal after etching is referred to as the specimen, and the direction from the boundary to the original substrate is defined as the positive direction.

Fig. 1

Appearance of the welding specimen. (a) Whole die and sample region, (b) sample region after polishing and etching.

2.2 Characterization

The hardness of the HAZ region was evaluated using an MMT-X3 Vickers hardness tester (Matsuzawa Corporation, Japan) with a load of 4.90 N. Hardness was tested over the range from −3.0 mm to +10.0 mm relative to the reference point at intervals of 0.2–1.0 mm. Microstructural images of the specimen surface were acquired using optical microscopy (OM; GX71, Olympus, Japan) and field-emission scanning electron microscopy (FESEM; JSM-7001F, JEOL, Japan) equipped with an energy dispersive spectrometry (EDS) system. Secondary electron images were obtained at an accelerating voltage of 10–15 kV and an emission current of 1.0 nA. The crystallographic structure was analyzed using X-ray diffraction (XRD) with an ULTIMA IV diffractometer (Rigaku, Japan). The primary X-ray source of Cu Kα radiation was focused with a spot diameter of 400 µm to obtain the information from a selected area. The diffraction angle was measured in the range of 30–120°. The OM, FESEM, and XRD analysis locations were selected based on the hardness analysis results.

3. Results

3.1 Hardness distribution

Figure 2 shows the hardness distribution in the HAZ and original substrate for the Mo-modified and standard H13 steel plates. Here, we examine only the hardness distribution of the Mo-modified steel plates. The results for the standard plate will be discussed later, together with the microstructural results. Regarding the Mo-modified H13 steel, the HAZ near the reference point (+0.2 mm) was the hardest and was approximately 140 HV harder than the original substrate. The hardness gradually decreased away from this point, yet it remained higher than that of the original substrate up to approximately +2.0 mm. Beyond +2.2 mm, the hardness dropped dramatically, being softer than the original substrate. The point at +2.8 mm shows the lowest hardness, which was approximately 200 HV lower than that of the original substrate. The hardness gradually recovered toward that of the original substrate.

Fig. 2

Hardness distribution in the welding specimen.

Based on the aforementioned results, structural analysis for the Mo-modified steel was performed at the following five points (Fig. 2): the most hardened point at +0.2 mm (HAZ (1)), the hardened region at +1.4 mm (HAZ (2)), the most softened point at +2.8 mm (HAZ (3)), the softened region at +4.0 mm (HAZ (4)), and the region corresponding to the base substrate without any heat affect at +8.0 mm (substrate). The denotations of “HAZ (X)” and “substrate” index the aforementioned analysis points.

3.2 Microstructural analysis for the modified steel

Figure 3 shows the OM images of the original substrate and HAZ for the Mo-modified H13 steel. Focusing on the substrate, the OM image corresponds to a perlite microstructure with grain sizes of approximately 10 µm. The image of HAZ (1) show coarsened grains several tens of micrometers in size (indicated by solid arrows), including elongated lath structures (indicated by dotted arrows). At HAZ (2), the compacted lath structure is included, and the grain size is refined to ∼5 µm. In both the HAZs (3) and (4) images, the grain sizes are apparently similar to that of the substrate—approximately 10 µm, and the perlite structure is also observed in HAZ (4).

Fig. 3

Optical microscope images of HAZs and substrate. (a) HAZ (1), (b) HAZ (2), (c) HAZ (3), (d) HAZ (4), and (e) Substrate.

The detailed microstructures, particularly the precipitates formation, were confirmed using FESEM (Fig. 4). The FESEM image of the substrate included the granular-like precipitates with a black contrast (indicated by solid arrow). The images of HAZ (1) shows the appearance of lath structures (indicated by dotted arrows) and plate-like precipitates (indicated by solid arrows). The lath structure also appears in the HAZ (2) image, although it is smaller in size. Further, granular precipitates and black precipitates (indicated by solid arrows) with diameters of ∼0.4 µm are also increased. The lath structure is not found and numerous fine granular precipitates with diameters of ∼0.1 µm (indicated by dotted arrows) are observed in the HAZ (3) image, which corresponds to an overtempered microstructure, comprising ferrite and cementite phases together with granular and plate-like precipitates [6, 13, 22]. The structure of HAZ (4) is similar to that of the substrate region, except that it contains the fine granular precipitates (indicated by solid arrows).

Fig. 4

FESEM images of HAZs and substrate. (a) HAZ (1), (b) HAZ (2), (c) HAZ (3), (d) HAZ (4), and (e) Substrate.

An EDS line analysis was conducted to identify the precipitates observed in the FESEM images (Fig. 5). The precipitates appear to be categorized into four types: the plate-like precipitates observed in HAZs (1) and (2), the fine granular precipitates of ∼0.1 µm in HAZs (3) and (4), the granular precipitates of ∼0.4 µm in HAZs (2), (3), and (4), and the back precipitates in HAZs (3), (4), and the substrate. The line analysis of the plate-like precipitates (along Line A in Fig. 4(a)) reveals the presence of high amounts of carbon without Mo or V, corresponding to Fe carbide. The black precipitate along Line B in Fig. 4(b) shows the presence of C, V, and Mo, while the granular precipitate along Line C shows C and Mo, and the fine granular precipitate along Line D in Fig. 4(c) shows only C. These results imply that the plate-like precipitates and the fine granular precipitates were Fe carbides, whereas the granular precipitates were Mo carbide. The back precipitates were complex carbide, including Fe, Mo, and V.

Fig. 5

EDS line analysis of the precipitates in Fig. 4. (a) Line A, (b) Line B, (c) Line C, and (d) Line D.

3.3 Crystallographic analysis for the Mo-modified H13 steel

Figure 6(a) shows a typical XRD pattern collected over a wide range of diffraction angles. The overall patterns are similar, irrespective of the location in the HAZ. Thus, we conducted detailed analysis focusing on the intense peak corresponding to ferritic Fe, α110. Figure 6(b) shows the enlargement of such α110 peaks collected from HAZ (1) and (2) in the hardened region. The pattern from the substrate is depicted for comparison. Diffraction angles of the α110 peaks for HAZs (1) and (2) exhibit a shift to lower angles compared with the substrate, and a tiny peak from austenitic Fe, γ111 is also found. The shift to lower angles originates from the increase in solid solution carbon in the bcc lattice structure, and the peak shift in HAZ (1) is larger than that in HAZ (2). The full width at half maximum (FWHM) spread originates from the lattice strain. These results imply that martensite was formed and that the carbon content in HAZ (1) was higher than that in HAZ (2) [9, 23, 24]. These results are consistent with the microstructural images shown in Figs. 3 and 4, where the lath structure is observed in the FESEM images from both HAZs (1) and (2), and such a structure is more prominent in HAZ (1).

Fig. 6

XRD patterns of the surfaces of the HAZ regions and substrates. (a) Wide angle diffraction pattern for the substrate. (b), (c) Enlargements around the α110 peak.

Figure 6(c) shows the enlarged XRD patterns from HAZs (3) and (4), which correspond to the softened region. The pattern from the substrate is also shown. The peak shapes are similar, and the diffraction angles of the α110 peaks in these HAZs are slightly shifted to higher angles compared to those of the substrate. Furthermore, the peak that corresponds to γ111 is not present in these HAZs. The origin of higher angle peak shift will be explained in the next section.

4. Discussions

4.1 Relationship of hardness and structural change in the HAZ

First, we confirm the substrate structure. The basic microstructure is similar to that of typical fine perlite, which is similar to that of tempered AISI H13 steel [5, 11, 12, 14, 15, 17, 18, 25, 26]. Furthermore, tiny precipitates, assumed to be carbides involving complex carbides as MC, M2C and M6C, were dispersed in the tempered structure; this dispersion enhances the hardness owing to the secondary hardening effect [13, 18, 19, 27].

Regarding the hardened region corresponding to HAZ (1) and (2), the lath structure, a typical microstructure of a martensite phase, is observed in Figs. 4(a) and (b). The occurrence of a martensitic transformation was also evidenced by the XRD patterns, specifically by the presence of the austenite peak γ111, spreading of the FWHM, and a shift of the peaks to lower angles. The spread and shift are caused by the increased lattice spacing and dislocation density, respectively (Fig. 6(b)). Thus, solute strengthening and higher dislocation density are the primary mechanisms underlying the formation of the hardened-region [9, 14, 18]. It was inferred that the temperature of this hardened region exceeded the upper critical temperature of the austenite transformation, AC3 at 1158 K [8, 9, 15, 18, 20]. The temperature during welding was certainly higher at HAZ (1) than that at HAZ (2). Additionally, the XRD peak shift (Fig. 6(b)) in HAZ (1) was larger than that in HAZ (2). On this basis, HAZ (2) was found to be composed of a lower content of the dissolved C in the martensite phase [18, 28]. Furthermore, the granular precipitates identified as Mo carbides were found in HAZ (2), whereas they were scarce in HAZ (1). V and Mo carbides in AISI H13 steel gradually dissolve as the temperature increases and would be completely dissolved above 1423 K [27, 29, 30].

The HAZ (3) region, corresponding to the most softened region, exhibited white needle-like structures in the intragranular region, which correspond to ferritic steel (Fig. 4(c)). Further, the austenitic Fe peak, γ111, was not present in the XRD pattern (Fig. 6(c)). These results indicate that the corresponding region was not beyond the lower critical temperature of the austenite transformation, AC1 at 1078 K [20, 28]. Moreover, the presence of Fe, V and/or Mo carbides precipitates in the corresponding FESEM images indicates the increment of these carbide precipitates. The XRD peak shift to higher angle is likely to be derived by the reduction of C in the intergranular region due to the increase of carbide precipitates [31]. This characteristic has been observed in typical overtempered structures, leading to the loss of the secondary hardening effect [12, 13, 1719, 25, 27, 28, 32, 33]. Consequently, the HAZ (3) region was significantly softened compared to the substrate region.

In HAZ (4), which was slightly softened compared to the substrate, the basic structure was almost the same as the substrate region, and the microstructural images were similar to the substrate. A few granular precipitates, corresponding to Fe, V and/or Mo carbides, were observed in the FESEM image (Fig. 5(d)), and the peak shift in XRD showed a similar tendency with the HAZ (3). The formation of such granular carbide precipitates leads to a decrease in hardness owing to the reduction of the tiny carbides, such as MC, M2C, and M6C, contributing to the secondary hardening effect [8, 1113, 17, 18, 25, 27, 30].

4.2 Relationship of temperature gradient and structural change in the HAZ

In this study, the HAZs are considered to be formed by heating the original Mo-modified H13 hot-work tool steel plate to varying temperatures that increase toward the boundary of the weld metal. This explanation is consistent with the observed tempered microstructure of the HAZ. The principle of the formation of the hardness distribution can be examined based on the temperature gradient in the HAZ.

Evidently, the highest temperature in the HAZ was observed at the boundary of the weld metal, which reached near the melting point, i.e., 1756 K [10]. The local maximum temperature decreases gradually with distance. Thus, HAZs (1) and (2) were likely heated more than AC3; hence, this region included martensite. Similar results regarding the enhancement of hardness in the boundary region of the weld metal have been reported [18, 25, 30, 34]. Moreover, the temperature in HAZ (1) was likely to have exceeded both the completion temperature of the austenite transformation, AC3 and the dissolution temperature of Mo and V carbides [18, 29, 30]. Thus, these carbides were dissolved in the lattice spacing, thereby enhancing both the solute and dislocation strengthening effects [18, 35].

By contrast, the temperature in HAZ (3) was lower than AC1 and thus austenitic transformation did not occur. In such cases, the microstructure exhibits a typical overtempered structure, and the hardness decreases with increasing temperature up to AC1 [1013, 18, 28]. Furthermore, the temperature in HAZ (4) only slightly exceeded 873 K, which was comparable to the tempering temperature during the preparation of the steel plate [21]. Thus, although the microstructure of the substrate region was essentially preserved, the quantities of Fe, Mo, and V carbides precipitates were increased, which led to a moderate decrease in hardness [11, 12, 19, 25, 27, 32].

4.3 Comparison of the Mo-modified steel with the standard H13 steel

Finally, we compare the Mo-modified steel with the standard H13 steel. The hardness distribution in the HAZ for the modified steel is similar to those of the standard H13 steels, yet the hardness values are different: the values in the hardened and the softened regions for the modified steel were lower than those for the standard steel (Fig. 1). The FESEM images for the hardened and softened regions are shown in Fig. 7. The fine granular precipitates (indicated by solid arrows) of ∼0.1 µm in sizes, and the black precipitates (indicated by dotted arrows) are found in the images, yet the granular precipitates of ∼0.4 µm are never found therein. The EDS line analysis evidenced that the fine granular precipitates agreed with the Fe carbides, and the black precipitates was with the complex carbide comprising V and Mo. Mo carbide was not found in the FESEM image. The Mo-modified steel includes a higher content of Mo, yet the concentration of C that contributes to the hardening of quenched steel is lower. Further, Mo carbides were hardly formed in the softened region, implying that the reduction of C in the intergranular region was mitigated. Hence, the hardness at the hardened and softened regions became lower than that of the standard steel.

Fig. 7

FESEM images of the standard H13 steel. (a) Hardened region of HAZ and (b) softened region of HAZ.

5. Conclusion

The microstructure of the HAZ in a welding-repaired Mo-modified AISI H13 hot-work tool steel are affected by the temperature during welding, with resultant changes in hardness. When the region of the HAZ near the boundary of the weld metal was raised above the austenite transformation temperature, thereby forming a martensite microstructure, and thus, the hardness increased notably. By contrast, the temperature in the region of the HAZ further away from the weld metal did not exceed the transformation temperature and instead only facilitated carbide precipitation. Thus, the corresponding area was softened compared to the original steel substrate owing to the loss of the secondary hardening effect. The formation of both hardened and softened regions in the HAZ in the Mo-modified H13 steel drives thermal fatigue crack reoccurrence; consequently, the lifetime of a repaired casting die is significantly shortened.

Acknowledgments

The authors are grateful to Prof. D. Kuroda of Suzuka of the National Institute of Technology, Suzuka College for providing his critical advice on the metallographic analysis. This study was supported by the North Forum of the Hokkaido branch of the Iron and Steel Institute of Japan.

REFERENCES
 
© 2025 The Japan Institute of Metals and Materials
feedback
Top