Behavior of Forging Steels under Cyclic Loading—the Benefit of Air‐Hardening Martensites

The development of air‐hardening martensitic forging (LHD: luft härtend duktil) steels offers high material performance with a short and simple process route. In this study, five alloys (L1–L5), based on the existing LHD alloy concept but with different contents of aluminum, titanium, boron, and molybdenum, are cast at laboratory scale. The casted blocks are hot forged into semifinished products and cooled in air (uncontrolled). The tensile properties, the Charpy V‐notch impact energy, the cyclic material behavior, and the fatigue strength of the alloys L1–L5 are opposed to each other. Furthermore, the material properties are compared with the standard quench and tempered (Q+T) steel 42CrMo4 (reference material) and ranked against previously developed forging steels. The tensile properties and Charpy V‐notch impact energy are comparable with those of the reference material, whereas the new alloy concepts show a significantly higher cyclic yield strength and fatigue strength.

DOI: 10.1002/srin.202000172 The development of air-hardening martensitic forging (LHD: luft härtend duktil) steels offers high material performance with a short and simple process route. In this study, five alloys (L1-L5), based on the existing LHD alloy concept but with different contents of aluminum, titanium, boron, and molybdenum, are cast at laboratory scale. The casted blocks are hot forged into semifinished products and cooled in air (uncontrolled). The tensile properties, the Charpy V-notch impact energy, the cyclic material behavior, and the fatigue strength of the alloys L1-L5 are opposed to each other. Furthermore, the material properties are compared with the standard quench and tempered (QþT) steel 42CrMo4 (reference material) and ranked against previously developed forging steels. The tensile properties and Charpy V-notch impact energy are comparable with those of the reference material, whereas the new alloy concepts show a significantly higher cyclic yield strength and fatigue strength. QþT steels. The carbide-free bainitic microstructure leads to high tensile properties and a high Charpy V-notch impact energy [7,[11][12][13] (see Figure 1b). However, because of the low phase transformation temperature of the bainite (Figure 1a), isothermal cooling after hot-forging is necessary to achieve a fully bainitic microstructure. The HDB alloy concept has been further modified to contain a certain amount of retained austenite as a second phase. This leads to the increased fatigue strength of the so-called transformation-induced plasticity (TRIP) steels, due to a deformation-induced transformation of austenite to martensite. [14][15][16] Recently, the focus has been set on simplifying cooling conditions while maintaining high tensile properties. Air-hardening (LHD: luft härtend duktil) steels offer high strength, comparable with QþT steels (Figure 1b), and an easy process route in common with PHFP steels. [17][18][19] By adding around 4 wt% manganese, a nearly fully martensitic microstructure is achieved by air cooling. [17][18][19] Unfortunately, manganese segregates to former austenite grain boundaries, leading to a low Charpy V-notch impact energy. [20] Alloying with boron [21,22] and/or molybdenum [23] seems to be a possible way to strengthen the grain boundaries and to increase the Charpy V-notch impact energy. Furthermore, decreasing the former austenite grain size is also an option for increasing the Charpy V-notch impact energy. [19]

Experimental Section
The new alloy concepts are compared with the QþT steel 42CrMo4, which is therefore set as the reference steel and represents the maximum alloying costs. The development of the LHD steels led to the alloy LHD-11, which contains around 4 wt% manganese, 0.2 wt% carbon, and 0.5 wt% silicon (see Table 1). This manganese content successfully suppresses the diffusion-controlled transformations, whereas higher manganese content would lead to embrittlement. [18,19] Silicon is alloyed to avoid cementite formation. [18,19] Based on this concept, for the new developments the alloying elements niobium, boron, molybdenum, titanium, and aluminum were alloyed in different amounts to achieve a high Charpy V-notch impact energy and increased fatigue properties, resulting in five alloy concepts ( Table 1). All alloys contain 0.035 wt% niobium for grain refinement of the prior austenite grain boundaries. For alloys L1 and L2, the amount of boron was varied to strengthen the grain boundaries and titanium content was adjusted to prevent the formation of boron nitrides ( Figure 2). Alloy L3 contains no significant amounts of boron and titanium but was alloyed with molybdenum for grain boundary strengthening ( Figure 2). L4 was alloyed with both boron and molybdenum; furthermore titanium was replaced by high amounts of aluminum ( Figure 2).  [7] amended by LHD steels. b) Charpy V-notch impact energy at room temperature versus yield strength for different forging steels. [8]  Alloy L5 contains no significant amounts of B, Ti, Mo, and Al. The carbon content of the alloys was also varied from 0.19 to 0.15 wt% (see Figure 3). For all five alloys, a block was ingot cast at the Steel Institute using a laboratory-scale vacuum furnace. The Steel Institute of RWTH Aachen University is able to reproduce the hot-forging manufacturing process with their Semi Product Simulation Center (SPSC). Using the SPSC, each block was austenitized for 5 h at 1200 C, then hot forged from 140 Â 140 Â 500 mm 3 to 60 Â 60 Â 1000 mm 3 in multiple steps and cooled in air (uncontrolled). The alloy LHD-11 was produced with the same procedure. For the QþT steel 42CrMo4, a bloc of 130 Â 130 Â 100 mm 3 was hardened for 72 min at 880 C, cooled in oil to room temperature, annealed for 167 min, and cooled in air to room temperature. [16] Samples for the characterization of material properties were taken from the transition between the center and edge of the blocks. Mechanical characterization was conducted through tensile tests according to DIN EN ISO 6892-1 [24] and Charpy-V-notch tests at room temperature according to DIN EN ISO 14556. [25] The cyclic material behavior was investigated through incremental step tests (ISTs). ISTs represent strain-controlled fatigue tests  with variable amplitudes on unnotched round specimens (K t ¼ 1) under alternating loading (R ε ¼ -1). The investigation of the fatigue strength in the low-cycle fatigue (LCF) and in the HCF regime was performed by load-controlled stress amplitude (SN, number of cycles to failure)-fatigue tests on notched specimens (K t ¼ 2) under tensile loading (R σ ¼ 0). The SN curves were evaluated with the maximum-likelihood method [26] and the fatigue strength was determined at N ¼ 1 Â 10 7 cycles with a 50% probability of survivability. The specimen geometries and the amount of samples for each alloy can be extracted from Table 2. Further information regarding the setup for tensile, Charpy V-notch, and cyclic testing has been published in Schmiedl and Gramlich. [27] Unloaded portions of the Charpy V-notch and notched fatigue specimens were used for microstructural analysis. Nital etching was used for examination of the microstructure. The surfaces of the fractured notched fatigue specimens were analyzed by scanning electron microscopy, using a ZEISS Supra 25 and secondary electron detector with an acceleration voltage of 20 kV.

Microstructure
Optical microscopy reveals a nearly fully martensitic microstructure for the alloys L1-L4 and a martensitic/bainitic microstructure for L5 (see Figure 4). The volume fraction of bainite increases, for all alloys, from the surface to the core of the forged block. The prior austenite grain sizes for alloys L1-L5 are 27, 50, 59, 46, and 40 μm, respectively. [28] Several types of precipitates were found in the alloys L1-L5; a detailed microstructural investigation has been reported in Gramlich et al. [28] All five alloys contain carbides as well as manganese sulfides. [28] In addition, alloys L2 and L3 contain large titanium nitrides and alloy L4 contains aluminum nitrides. [28]

Tensile Properties
The reference steel 42CrMo4 shows a yield strength of 954 MPa and a tensile strength of 1090 MPa ( Figure 5). The former LHD steel concept, LHD-11, has already led to a comparable yield strength (909 MPa) and to an increase by 430 MPa in tensile strength (1523 MPa), in comparison with the reference steel 42CrMo4 ( Figure 5). The alloys L1-L5 show a yield strength in the range of 918-993 MPa and a tensile strength in the range 1340-1511 MPa ( Figure 5). Both strength values decrease from alloy L1 to L4 and then increase for alloy L5. The alloy L5 shows almost an identical yield and tensile strength to L2 ( Figure 5). The tensile properties of the new alloys are in the range of the alloy LHD-11, so the main advantage is a higher tensile strength in comparison with the reference steel 42CrMo4.
With reference to the ductility, the reference steel 42CrMo4 and the alloy LHD-11 show a uniform elongation of about 5.0% (see Figure 6). The alloys L1-L5 show a lower uniform elongation by around 1.0% ( Figure 6). The elongations to fracture of the alloys L1-L5 and of the alloys LHD-11 are about 0.5% lower in comparison with the reference steel 42CrMo4 ( Figure 6). The new alloys offer a higher performance than the QþT steel 42CrMo4 with respect to the tensile properties and the process route.

Charpy V-Notch Impact Energy at Room Temperature
The former LHD steel, LHD-11, reaches only a moderate Charpy V-notch impact energy of 29 J, whereas the reference steel 42CrMo4 shows a high Charpy V-notch impact energy of 51 J, see Figure 7. The Charpy V-notch impact energies of the alloys L1-L5 show a completely different trend than the tensile properties. Only alloy L4 shows a high Charpy V-notch impact energy comparable to the reference steel 42CrMo4 (Figure 7). The alloys L1, L3, and L5 show a relatively low Charpy V-notch impact energy of 11-16 J and L2 shows a moderate Charpy V-notch impact energy of 25 J (Figure 7).

Cyclic Material Behavior
ISTs provide the monotonic yield strength from the initial loading curve and the cyclic yield strength from the cyclic   [16] the LHD steel LHD-11, [19] and the alloys L1-L5. [27] Figure 6. Uniform elongations A g and elongations to fracture A 5 of the QþT steel 42CrMo4, [16] the LHD steel LHD-11, [19] and the alloys L1-L5. [27] Figure 7. Charpy V-notch impact energies at room temperature of the QþT steel 42CrMo4, [16] the LHD steel LHD-11, [19] and the alloys L1-L5. [27] www.advancedsciencenews.com stress-strain curve. Initial loading and cyclic stress-strain curves of the alloys L1-L5 have been shown elsewhere. [27] The comparison between the monotonic and the cyclic yield strength indicates the cyclic material behavior. If the cyclic yield strength is higher than the monotonic yield strength, cyclic hardening occurs and, if it is lower, cyclic softening takes place. A cyclically neutral material behavior appears if no significant difference exists between monotonic and cyclic yield strengths.
The standard QþT steel 42CrMo4 shows a cyclic softening leading to a low cyclic yield strength of 560 MPa (see Figure 8). In contrast, the alloys L1-L5 show a cyclic hardening and the cyclic yield strengths of the alloys L1-L5 are about 70-90% higher than that of the QþT steel 42CrMo4 (Figure 8). Analogous to the tensile strength, the cyclic yield strengths of the alloys decrease from L1 to L4 and then increase for L5 (Figure 8).

Fatigue Strength of Notched Specimen in the HCF Regime
Spring clips, axles, and stub axles are typical forging components that are designed for long lifetime and are often loaded with tensile mean stresses. Therefore, a high fatigue strength in the HCF regime is necessary to perform a lightweight design. The QþT steel 42CrMo4 shows a fatigue strength in the HCF regime of 194 MPa (N ¼ 1 Â 10 7 cycles) (see Figure 9). The fatigue strengths in the HCF regime (N ¼ 1 Â 10 7 cycles) of all the new alloys (L1-L5) are about 25-48% higher in comparison with the reference steel 42CrMo4 (Figure 9). The alloys L1 and L2, as well as L3-L5, show a comparable fatigue strength of 245 MPa and 280 MPa, respectively. The alloy L4 offers the best performance with a fatigue strength of 286 MPa (Figure 9).

Fatigue Fracture Behavior
Fractographic analyses of the notched fatigue specimens were performed to identify the influence of the alloying elements on the crack initiation and propagation behavior. Therefore, specimens with cycles to failure close to the knee point were analyzed for the alloys L1-L5. The alloys L1 and L2 show a transcrystalline fatigue fracture and the residual fracture is characterized by a transcrystalline overload fracture with dimples (see Figure 10). Precipitates are present at the crack initiation location for the alloys L1 and L2 (Figure 10). Transcrystalline and intercrystalline fatigue fracture takes place for the alloys L3 and L5 ( Figure 10). No precipitates as crack initiation point could be determined at the origin of fracture. The residual fracture shows an intercrystalline fracture behavior and the amount of dimples is less pronounced for L5 in comparison with L3 ( Figure 10). In contrast, alloy L4 shows a purely transcrystalline fatigue fracture and a purely ductile residual fracture ( Figure 10). Similar to L3 and L5, the origin of fracture is not located at precipitates.

Discussion
The tensile properties of the alloys L1-L5 indicate that tensile strength mainly depends on the carbon content ( Figure 5). Increasing carbon content leads to a higher tensile strength. Microalloying elements and grain size seem to play a minor role on the tensile properties of the alloys L1-L5.
In contrast, the Charpy V-notch impact energy seems to depend mainly on the microalloying elements. Alloy L4, which is alloyed with both grain boundary strengthening elements boron and molybdenum, shows the highest Charpy V-notch impact energy of 50 J (Figure 7). Comparison of the alloys L3 and L5 shows that molybdenum effectively increases the impact energy but these alloys still show a relatively low impact energy in comparison with the boron-containing alloys (LHD-11, L1, L2). In the case of the boron-containing alloys, it is necessary that the boron is not bonded to other elements and can diffuse to the grain boundaries for strengthening. Therefore, nitride-forming elements, such as titanium and aluminum, are alloyed, consuming the nitrogen and forming nitrides to keep the boron free. [29][30][31] Figure 8. Monotonic yield strengths from initial loading curve and cyclic yield strengths R* p0.2 from IST of the QþT steel 42CrMo4, [16] the LHD steel LHD-11, [19] and the alloys L1-L5. [27] Figure 9. Fatigue strength of notched specimens in the HCF regime of the QþT steel 42CrMo4, [16] the LHD steel LHD-11, [19] and the alloys L1-L5. [27] www.advancedsciencenews.com l www.steel-research.de steel research int. 2020, 91, 2000172 Therefore, higher boron contents require also the alloying of higher amounts of titanium or aluminum. The alloy LHD-11, with 0.0029 wt% boron, shows a higher Charpy V-notch impact energy than L1, with a boron content of 0.0016 wt% (Figure 7). This leads to the assumption that the Charpy V-notch impact energy increases with increasing boron content. However, alloy L2, which contains 0.0057 wt% boron, shows a lower impact energy than LHD-11. The reason for this is that the titanium content is too low to prevent the formation of boron nitrides, which results in minor grain boundary strengthening and therefore in a lower Charpy V-notch impact energy. In contrast, alloy L4 is hyperstochiometrically alloyed with aluminum to completely ensure that boron nitride formation is not possible. In addition, investigations reported in the literature have shown that around 0.0030 wt% boron gives a saturation level of grain boundary strengthening, [29] although this cannot be rebutted or confirmed by this study. Nevertheless, boron effectively strengthens the grain boundaries and increases the impact energy. All in all, the Mo-B-Al concept (L4) leads to the best combination of high strength and high Charpy V-notch impact  energy of all the new alloys. In comparison with other forging steels, the Mo-B-Al LHD concept (L4) offers, with a short and simple process route, an outstanding performance regarding high tensile properties and high Charpy V-notch impact energy ( Figure 11). The cyclic material behavior and the fatigue strength also play an important role in performing a lightweight design. Despite high tensile properties, the reference steel 42CrMo4 shows a low cyclic yield strength of 560 MPa due to cyclic softening (see Figure 8). In contrast, the five alloys L1-L5 show cyclic hardening leading to a high cyclic yield strength of around 1000 MPa (Figure 8). In general, QþT steels show cyclic softening, PHFP steels shows a cyclically neutral behavior, and TRIP steels, as well as the new LHD alloys, show cyclic hardening (see Figure 12a). The values of the cyclic yield strength of the alloys L1-L5 follow a similar trend to the strength values of tensile properties. This indicates a correlation between the cyclic yield strength and the tensile properties. For the alloys L1-L5, it is not clear whether the cyclic yield strength mainly depends Figure 11. a) Elongation to fracture A 5 versus tensile strength R m and b) Charpy V-notch impact energy at room temperature versus yield strength R p0.2 for different forging steels (PHFP, [4,5] HDB, [7] TRIP, [16] LHD, [19] and QþT [4,5,16] ) and for the new LHD-alloys. [27] Figure 12. Cyclic yield strength from IST versus a) monotonic yield strength from initial loading curve of IST, b) the yield strength from tensile tests, and c) the tensile strength from tensile tests for different forging steels (PHFP, [4,5] TRIP, [16] LHD, [19] and QþT [4,5,16] ) and the new LHD alloys. [27] www.advancedsciencenews.com on the yield or on the tensile strength. However, investigations of other forging steels lead to the assumption that the cyclic yield strength correlates well with the tensile strength and less well with the yield strength (see Figure 12b,c). The alloys L1-L5 also show improved fatigue strength in the HCF regime in comparison with the reference steel 42CrMo4 (Figure 9). No correlation of the fatigue strength in the HCF regime with the tensile properties or the cyclic yield strength could be determined for the alloys L1-L5. However, the alloys that are alloyed with titanium (L1, L2) show a lower fatigue strength in the HCF regime than the alloys without titanium (L3-L5) (see Figure 9). Titanium nitrides are observed in the alloys L1 and L2 and are characterized by a sharp morphology, as reported elsewhere. [28] Therefore, they act as inner notches which favor crack formation under cyclic loading. For alloy L4, small and round aluminum nitrides are observed, [28] but these are not as critical for fatigue crack formation as titanium nitrides due to their morphology. In general, the fatigue strength depends on the tensile properties of a material (yield strength and tensile strength), as shown in Figure 13. However, in detail, the fatigue strength is strongly affected by microstructural defects. [32] The surface fractures of the notched fatigue specimens of the alloys L1 and L2 indicate that boron effectively strengthens the grain boundaries and prevents intercrystalline fracture (see Figure 10). Due to the significant fraction of intercrystalline fractures for the molybdenum-containing alloys L3 and L5, the effect of molybdenum on the grain boundary strengthening is less pronounced (Figure 10). The addition of 0.5 wt% aluminum in combination with boron and molybdenum (L4) leads to a completely transcrystalline fracture with a ductile residual fracture ( Figure 10). A comparable trend, for the alloys L1-L5, is also observed for the fracture behavior of the Charpy V-notch impact energy specimens. [28] Crack initiation takes place at precipitates for the alloys L1 and L2 (Figure 10), explaining the lower reported fatigue strength of these alloys in comparison with the alloys L3-L5. Gramlich et al. [28] showed that titanium nitrides possess a rectangular morphology with sharp edges, whereas aluminum nitrides appear round and cylindrical. Therefore, titanium nitrides seem to favor fatigue crack initiation, leading to a lower fatigue strength.

Conclusions
The alloy group of air-hardening martensites offers the best combination of tensile and cyclic properties of the presented forging steels groups (PHFP, HDB, TRIP, QþT). Further advantages are low alloying, processing and machining costs as well as low CO 2 -emissions because additional heat treatments are not necessary.
The following conclusions can be drawn from this study: 1) The cyclic yield strengths correlate with the tensile properties, especially with the tensile strength, which itself depends on the carbon concentration dissolved in the matrix. 2) The fatigue strength in the HCF regime is connected to the size and morphology of the nitrides and is therefore influenced by the nitride-forming elements titanium and aluminum. Titanium nitrides in contrast to aluminum nitrides act as inner notches, negatively affecting the fatigue strength.
3) The prior austenite grain boundary strengthening is also increased under cyclic loading by boron and molybdenum, but the effect of boron is more pronounced. 4) The Mo-B-Al LHD concept offers the best combination of high tensile properties, high impact energy, and high cyclic properties. Figure 13. Fatigue strength of notched specimens in the HCF regime versus a) the yield strength from tensile test and b) the tensile strength from tensile tests for different forging steels (PHFP, [4,5] TRIP, [16] LHD, [19] and QþT [4,5,16] ) and for the new alloys. [27] www.advancedsciencenews.com l www.steel-research.de steel research int. 2020, 91, 2000172