Hot Processing Maps and Texture Evolution during Hot Compression of CF170 Maraging Stainless Steel

Abstract

In this work, the hot processing maps of CF170 maraging stainless steel were generated by using flow stress data obtained from an isothermal compression experiment on the Gleeble-3800 hot compression simulator. The microstructure corresponding to the stable regions in the hot processing map was observed with an optical microscope. The results showed that the steel is susceptible to deformation temperature, strain, and strain rate, and the optimal hot working parameter is a strain rate range of 0.01∼0.1 s⁻¹ and a temperature range of 1350∼1373 K when the strain is 0.6. In addition, the texture evolution of prior austenite during hot compression was investigated based on reconstructing a parent grain by Electron Back-Scattered Diffraction. The results showed that twinning-induced dynamic recrystallization is another mechanism except for strain-induced dynamic recrystallization at a higher strain rate. Further, the deformed microstructure with a Cube texture (1 0 0) ⟨0 1 0⟩ will preferentially nucleate, and the increase in temperature makes the Cube texture (1 0 0) ⟨0 1 0⟩ change to the rotated Cube texture (1 0 0) ⟨1 1 0⟩, and the increase of strain rate inhibits the growth of recrystallized grains and weakens texture.

1. Introduction

Due to its corrosion resistance and excellent mechanical properties, high-strength stainless steel is widely used in many engineering fields.^1,2) However, some critical structural parts, such as landing gear, require not only high strength and excellent corrosion resistance but also high toughness. Therefore, maraging stainless steel with the advantages of maraging steel and precipitation hardening stainless steel is increasingly favoured. In recent years, some studies on maraging stainless steel containing Co focused on the effects of alloying elements on phase precipitation,³⁾ mechanical properties,^4,5) and corrosion resistance.⁶⁾ Compared with maraging stainless steel containing Co, CF170, new Co-free maraging stainless steel, developed by Central Iron & Steel Research Institute based on maraging steel, not only saves resources of Co but also has high strength and excellent toughness. CF170 maraging stainless steel has excellent high-temperature deformation resistance because of adding a large quantity of alloying elements, which makes its hot deformation behavior differ from that of traditional carbon steel and low alloy steel.

The high-temperature deformation and dynamic recrystallization behavior of CF170 steel have been investigated in our previous work.⁷⁾ However, there are still some bottlenecks in homogenization control for the large section forging process of CF170 steel, especially in the upsetting and drawing process of the traditional forging, which has a bit of blindness in the selection of parameters such as forging temperature forging speed and forging ratio. Therefore, the reasonable selection of forging parameters is very significant for achieving the homogeneous and refined microstructure after forging, which is conducive to further improving its toughness on the premise of ensuring the strength of this steel. Since Prasad et al.⁸⁾ proposed the processing map based on the dynamic materials model (DMM) as an effective means to optimize material processing parameters, it has been widely used to evaluate the thermal deformation behaviour of a variety of metal materials, including aluminium,⁹⁾ magnesium,¹⁰⁾ titanium,¹¹⁾ Co-based,¹²⁾ Ni-based alloys,¹³⁾ and stainless steel.¹⁴⁾ Chen et al.¹⁵⁾ investigated the processing maps of Mn18Cr18N austenitic stainless steel, which shows that microstructure is more refined in the domains with higher efficiency of power dissipation. That is, processing maps can reflect the materials’ dynamic recrystallization (DRX) degree. In general, the textures will evolve during the progress of recrystallization. Typical textures with Brass components and Cube/rotated Cube components were observed during the hot deformation of AISI 304LN austenitic stainless steel.¹⁶⁾ So far, much research has focused on texture evolution in alloys without phase transformation.^10,17) Nevertheless, texture evolution of high-temperature austenite in maraging stainless steel during hot deformation is rarely reported.

The present work aims to investigate the hot deformation behaviour and texture evolution during hot compression of the CF170 maraging stainless steel combined with the processing map and provide optimal deformation parameters for this steel’s large section forging process.

2. Experiment Material and Methods

Chemical composition (mass%) of the CF170 maraging stainless steel is as follows: ≤0.01 C, 10.5∼12.5 Cr, 10.7∼11.3 Ni, 0.75∼1.25 Mo, 1.2∼1.4 Ti and balance Fe. The hot compression samples were sampled from an ingot melted by vacuum self-consuming arc melting. Then these samples were machined into a cylinder with an initial dimension of Φ 8 × 15 mm. An isothermal compression experiment was performed on the Gleeble-3800 hot compression simulator at temperatures from 1173 to 1473 K under a strain rate range of 0.01∼10 s⁻¹. The samples were first heated to 1523 K for 600 s at a rate of 10 K/s to eliminate segregation, then cooled to the preset deformation temperature at a rate of 10 K/s and kept for 30 s to remove the temperature gradient. In order to retain the high-temperature austenite microstructure, all specimens were water quenched immediately after compressing to a true strain of 0.6. The experimental schematic is shown in Fig. 1. The quenched samples were sectioned along the axial direction, polished, and then electrolyzed at a voltage of 1.0∼1.2 V in nitric acid to observe the metallographic structure by optical microscope (OM). The texture evolution of high-temperature austenite was investigated based on reconstructing a parent grain by the EBSD (Electron Back-Scattered Diffraction). Samples for the EBSD scan were firstly mechanical polished and then electropolished at 14 V for 18 s in a perchloric acid alcohol solution (12.5% HClO₄ + 87.5% C₂H₆O, volume%). In order to facilitate the expression of crystal orientation, rolling direction (RD), transverse direction (TD), and normal direction (ND) of the samples are defined, respectively, as shown in Fig. 1. The RD of the samples is parallel to the X direction of the sample stage in EBSD test process. Therefore, the coordinate system needs to be rotated 90° around the RD when analyzing textures.

Fig. 1

Schematic of hot compression test.

3. Results and Discussion

3.1 Flow stress curves

Figure 2 shows the flow stress curves of CF170 maraging stainless steel in various deformation conditions. The stress decreases with the increase of temperature and increases with the increase of strain rate. On the one hand, the higher the strain rate, the higher the dislocation density. On the other hand, the increase in temperature will promote vacancy diffusion, dislocation slip and grain boundary movement.^18,19) In addition, the stress curves show two characteristics: DRX-type curves with a single peak and dynamic recovery (DRV) type curves without a single peak. The former represents that when the stress reaches the peak, it gradually decreases with the increase of the strain and finally gates a plateau, such as stress curves at a strain rate of 0.01 s⁻¹ and 0.1 s⁻¹ in Fig. 2(b). In contrast, DRV-type curves show that as the strain increases, the stress gradually increases until it reaches a stable stage, such as those stress curves at a low temperature or high strain rate. The reason for DRV-type stress curves at a high strain rate and high temperature is that there is not enough time for nucleation and growth of DRX grains, according to the Ref. 18). However, the stress curves of 17-4 PH stainless steel²⁰⁾ at a strain rate of 1∼10 s⁻¹ also show the DRV-type characteristic, but DRX still occurs, indicating the softening effect caused by DRX can not completely offset the work hardening during deformation.

Fig. 2

The flow stress curves in various conditions: (a) $\dot{\varepsilon } = 0.1$ s⁻¹, (b) T = 1473 K.

3.2 Processing maps and evolution of microstructure

Flow stress data are used to design the processing map based on the theory of DMM.²¹⁾ The hot deformation process of materials can be characterized by the total dissipated power (P) consisting of two parts of J and G. The co-content J relates to the metallurgical mechanism to dissipate power such as DRX, DRV, cavity formation, super-plastic flow, and phase transformation.²²⁾ The content G relates to the power dissipated in the form of temperature rise during deformation. Therefore, the P can be described as follows:⁸⁾   

\begin{equation} \text{P} = \text{J} + \text{G} = \int_{0}^{\sigma} \dot{\varepsilon}\,\text{d}\sigma + \int_{0}^{\dot{\varepsilon}} \sigma\,\text{d}\dot{\varepsilon} \end{equation} (1)

If the stress versus strain rate obeys a power law:   

\begin{equation} \sigma = \text{K}\dot{\varepsilon}^{\text{m}} \end{equation} (2)

Where K is the material constant, m is the exponent independent of the strain rate and calculated by eq. (3).⁸⁾   

\begin{equation} \text{m} = \frac{\partial\text{J}}{\partial\text{G}} = \frac{\partial (\ln \sigma)}{\partial (\ln \dot{\varepsilon})} \end{equation} (3)

The relationship between stress and strain rate can be expressed through a cubic polynomial for a given strain and deformation temperature:²³⁾   

\begin{equation} \ln\sigma = \text{a} + \text{b} \ln \dot{\varepsilon} + \text{c} (\ln \dot{\varepsilon})^{2} + \text{d} (\ln{\dot{\varepsilon}})^{3} \end{equation} (4)

According to eq. (3) and (4), the exponent m can be expressed as:   

\begin{equation} \text{m} = \text{b} + \text{2c} \ln \dot{\varepsilon} + \text{3d} (\ln \dot{\varepsilon})^{2} \end{equation} (5)

The efficiency of power dissipation η is defined as:⁸⁾   

\begin{equation} \eta = \frac{\text{J}}{\text{J}_{\text{max}}} = \frac{\text{2m}}{1 + \text{m}} \end{equation} (6)

where the maximum co-content $\text{J}_{\max } = \frac{\sigma \dot{\varepsilon }}{2}$ when m = 1.

The instability parameter ξ($\dot{\varepsilon }$) can be expressed as:²⁴⁾   

\begin{equation} \xi (\dot{\varepsilon}) = \frac{\partial \ln[\text{m}/(1 + \text{m})]}{\partial \ln \dot{\varepsilon}} + \text{m} < 0 \end{equation} (7)

The parameter ξ($\dot{\varepsilon }$) is used to plot an instability map, which indicates an unstable zone of deformation when the ξ($\dot{\varepsilon }$) is negative.

The cubic polynomial curves of ln σ versus ln $\dot{\varepsilon }$ can be obtained by fitting flow stress data of different deformation conditions at strains of 0.1, 0.2, 0.4, and 0.6, as shown in Fig. 3. Once the values of b, c, and d are determined, the value of m can be obtained. Then the values of ξ($\dot{\varepsilon }$) and η can be obtained according to eq. (3) and (4), respectively. Finally, the hot processing map can be plotted by superimposing the instability map on the power dissipation map.

Fig. 3

Linear fitting of ln σ versus ln $\dot{\varepsilon }$ at various strains: (a) 0.1, (b) 0.2, (c) 0.4, (d) 0.6.

The 3D surfaces of m at different true strains are shown in Fig. 4, in which there is no apparent regulation between the value of m and deformation temperature and strain rate. It is noticed that the values of m are negative at some higher strain rate regions, such as the strain rate of 10 s⁻¹ shown in Fig. 4(a)∼(d), which is caused by dynamic strain aging,²⁵⁾ deformation twinning, shear band formation, initiation, and growth of microcracks.^24,26)

Fig. 4

The 3D surfaces of m at various strains: (a) 0.1, (b) 0.2, (c) 0.4, (d) 0.6.

Figure 5 shows the variation in the efficiency of power dissipation η with the strain and temperature. It can be seen that the effect of deformation temperature on the efficiency of power dissipation is different. The values of η at a strain rate of 0.01 s⁻¹ with different temperatures show the following characteristics: i) the value of η rises first and then reaches a steady state when the temperature is 1173 K, which indicates DRV is a main softening mechanism under this deformation condition. The value of η rises first and then decreases when the temperature is 1273 K and 1473 K. However, the two are slightly different. The value of η at a temperature of 1273 K has hardly changed when the strain exceeds 0.4, which indicates partial DRX or DRV may be a main deformation mechanism at the current temperature. The value of η decreases when the temperature is 1473 K, which is caused by grain growth. ii) the value of η increases when the temperature is 1273 K and 1473 K, which indicates the DRX volume fraction increases continuously. The values of η at a strain rate of 10 s⁻¹ decrease with the strain increase due to the unstable flow bands.²⁷⁾

Fig. 5

Efficiency of power dissipation at various temperatures: (a) 1173 K, (b) 1273 K, (c) 1473 K.

The processing maps of CF170 steel with different strains are shown in Fig. 6. The shaded domains represent the instability regions of hot working, and the contour lines represent the efficiency of power dissipation η. As the strain increases, the area of the instability regions gradually decreases. When the strain rate is in the range of 0.01∼0.1 s⁻¹, the stability regions are divided into four regions marked as A, B, C, and D with different dissipation efficiency, as shown in Fig. 6(d). The values of η in regions B and D are greater than 0.3, while less than 0.25 in regions A and C.

Fig. 6

Processing maps of CF170 steel at various strains: (a) 0.1, (b) 0.2, (c) 0.4, (d) 0.6.

Region A occurs at the temperature range of 1273∼1310 K with a strain rate range of 0.01∼0.025 s⁻¹. Figure 7(a) shows the microstructure in the region with an efficiency of power dissipation of 0.228 at the temperature of 1273 K and a strain rate of 0.01 s⁻¹. Coase deformed grains are elongated perpendicular to the compression direction, and some small recrystallized grains along the grain boundaries present a necklace structure. In addition, some serrated grain boundaries were observed, which indicates that DRX has occurred in this deformation condition. Region B occurs at the temperature range of 1323∼1373 K with a strain rate range of 0.01∼0.15 s⁻¹. When the temperature is 1373 K and the strain rate is 0.01 s⁻¹, a large number of recrystallized grains are clearly observed, as shown in Fig. 7(b). At this time, the peak efficiency of power dissipation is 0.32, which does not reach the maximum value ∼0.4 of region D. It may be that DRX is incomplete under this deformation condition. The main reason is that the mobility of grain boundaries decreases with the decrease in temperature, which reduces the rate of forming new DRX grains by bulging grain boundaries. However, the efficiency of power dissipation in region C is less than 0.2 when the temperature ranges 1410∼1473 K with a strain rate range of 0.01∼0.02 s⁻¹. As mentioned earlier, grain growth results in a decrease in the efficiency of power dissipation. Figure 7(c) shows the microstructure at the temperature of 1473 K and the strain rate of 0.01 s⁻¹, in which the grain boundaries become tortuous, and the grains grow obviously. It indicates that recrystallized grains continue to grow with the increase in temperature. The efficiency of power dissipation is ∼0.4 at a temperature of 1473 K with a strain rate range of 0.1 s⁻¹. The corresponding microstructure presents refined equiaxed grains, as shown in Fig. 7(d). Compared with Fig. 7(c), the grain size decreases with the increase of the strain rate due to the decrease in grain boundary mobility.²⁰⁾

Fig. 7

Microstructure in various deformation conditions: (a) 1273 K–0.01 s⁻¹, (b) 1373 K–0.01 s⁻¹, (c) 1473 K–0.01 s⁻¹, (d) 1473 K–0.1 s⁻¹.

It is worth noting that η = −0.15 at 1473 K with a strain rate of 10 s⁻¹. The efficiency of power dissipation is negative, which results from the values of m are negative, according to above and eq. (15). Figure 8(a) shows microstructure at a temperature of 1473 K and a strain rate of 10 s⁻¹, in which refined equiaxed grains are observed, indicating DRX still occurred at a higher strain rate. Combined with the reconstructed parent phase in Fig. 8(c), a prior austenite grain outlined by the blue line consists of packets outlined by the yellow lines. As shown in Fig. 8(b), those packets are divided into some of the same colour parallel blocks. The misorientation angle of block boundaries is a range of 50∼70°, which is calculated based on the ideal KS orientation relationship.^28,29) Interestedly, large quantities of deformed twins marked with red lines are observed in Fig. 8(d), which is calculated to be a classic ⟨1 1 1⟩ 60° twin boundaries in FCC metals based on EBSD data processing. In other words, at a higher strain rate, in addition to nucleation at grain boundaries, twinning promotes nucleation to induce recrystallization is another recrystallization mechanism. DRX has occurred at a strain rate of 10 s⁻¹, but the flow stress curve still presents a DRV-type characteristic, as shown in Fig. 2(b), which may be related to twinning promoting nucleation. On the one hand, dislocations continuously accumulate at the twin boundaries with increased strain. On the other hand, the dislocation density increases with the strain rate. Consequently, the softening effect caused by DRX can hardly offset the hardening effect caused by dislocations stacking, resulting in DRV-type flow stress curves without a single peak.

Fig. 8

Microstructure and results of EBSD scanning at 1473 K–10 s⁻¹: (a) metallographic image, (b) EBSD misorientation mapping, (c) reconstructing parent grain, (d) EBSD grain boundary mapping.

Figure 9 shows the schematics of the recrystallized microstructure evolution in different deformation conditions. Due to the inhomogeneous strain distribution, the local dislocation density on both sides of the high-angle grain boundaries (HAGBs) is different, which makes the nucleation by bulging grain boundaries, namely, strain-induced grain boundary migration nucleation, finally forming the first layer of necklace-like grains (in blue) as shown in Fig. 9(a). Subsequently, grains continuously nucleate toward the interior of the deformed microstructure, gradually forming the next layer of necklace-like grains. At the same time, the work hardening occurs in the hot compression process, reducing grain growth’s driving force gradually. In addition, grain growth is also inhibited by the nucleation of the next layer of recrystallized grains. Therefore, the recrystallized grains size is small at low temperature and low strain rate. With the increase of strain rate, however, dislocations’ slip is difficult to fully carry out and disappear at grain boundaries in a short time, which leads to local stress concentration, thus promoting twinning.³⁰⁾ The dislocation density continues to increase with deformation; on the one hand, a large quantity of low-angle grain boundaries (LAGBs) (in grey) are formed; on the other hand, DRX occurs at twin boundaries (TBs) (in green) with the high storage energy, as shown in Fig. 9(b), namely, twinning-induced DRX. In addition, the migration speed of grain boundaries increases with the deformation temperature. It provides conditions for recrystallized grains growth to reduce interface energy.

Fig. 9

Schematics of the recrystallized microstructure evolution in various deformation conditions: (a) 1273 K–0.01 s⁻¹, (b) 1473 K–10 s⁻¹.

3.3 Texture evolution of prior austenite during hot compression

According to the above discussion, inhomogeneous strain distribution leads to layered necklace-like recrystallized grains. However, the strain in the central microregion of the sample is almost the same during the hot compression process, but the recrystallized microstructure is obviously different. As shown in Fig. 7(a), some necklace-like recrystallized grains only exist at the grain boundaries, while others completely replace the original deformed microstructure, indicating local orientation gradient in deformed grains plays an essential role in promoting recrystallization nucleation. Unfortunately, prior austenite was calculated based on reconstructing a parent grain, which makes it difficult to analyze the misorientation development in the deformed microstructure via the point-to-point/origin misorientation. Therefore, to further study the recrystallization nucleation orientation, Fig. 10 shows the IPF map and pole figures of the local recrystallized grains. The results show that almost all recrystallized grains form a stronger Cube texture (1 0 0) ⟨0 1 0⟩. A slight difference is that the Cube texture rotates slightly around the TD direction in the second and third group of polar figures, as shown in Fig. 10(b). In addition to twinning nucleation, new nucleus orientation will not be generated before and after the nucleation of recrystallized grains, so the deformed microstructure with a Cube texture (1 0 0) ⟨0 1 0⟩ will preferentially nucleate during the recrystallization.

Fig. 10

Results of reconstructing a parent grain at 1273 K–0.01 s⁻¹: (a) EBSD IPF map, (b) pole figures of different selected areas.

Figure 11 shows pole figures of prior austenite in various deformation conditions. As shown in Fig. 11(a)∼11(b), when the strain rate is 0.01 s⁻¹, the maximum random distribution (MRD) increases from 8.84 to 13.41 with the increase in temperature, indicating texture strength increases due to the rise of DRX volume fraction and growth of recrystallized grain when the temperature increases. Further, Fig. 11(a) shows that there is Cube texture (1 0 0) ⟨0 1 0⟩ at a temperature of 1373 K and strain rate of 0.01 s⁻¹. Cube texture rotates 45° around the TD direction to form rotated Cube texture (1 0 0) ⟨1 1 0⟩ when the temperature rises from 1373 K to 1473 K, as shown in Fig. 11(b). In other words, the rise of temperature increases the DRX volume fraction, and leads to preferential growth of recrystallized grains and the rotation combination of subgrains. Compared with Fig. 11(b), there is no noticeable texture in Fig. 11(c), which shows that the MRD reduces from 13.41 to 3.97, indicating the increase in strain rate can inhibit the preferential growth of recrystallized grains and weaken texture.

Fig. 11

Pole figures of prior austenite in various deformation conditions: (a) 1373 K–0.01 s⁻¹, (b) 1473 K–0.01 s⁻¹, (c) 1473 K–10 s⁻¹.

4. Conclusions

Hot processing maps and texture evolution of CF170 maraging stainless steel were studied by isothermal compression experiment in the temperature range of 1173∼1473 K and the strain rate range of 0.01∼10 s⁻¹. The main conclusions of this research are obtained as follows:

1. (1)    CF170 steel is susceptible to deformation temperature, strain, and strain rate. When the strain is 0.6, the optimal deformation condition for hot working is a strain rate range of 0.01∼0.1 s⁻¹ and a temperature range of 1350∼1373 K.
2. (2)    DRX mechanisms are different for different deformation conditions in the compression process. Strain-induced grain boundary migration nucleation is the main DRX mechanism at a lower strain rate, while the twinning-induced DRX is another mechanism except for strain-induced DRX at a higher strain rate.
3. (3)    DRX process shows a strong nucleation orientation, and the deformed microstructure with a Cube texture (1 0 0) ⟨0 1 0⟩ will preferentially nucleate. The increase in temperature makes the Cube texture (1 0 0) ⟨0 1 0⟩ change to the rotated Cube texture (1 0 0) ⟨1 1 0⟩ due to the preferential growth of recrystallized grains and rotation combination of subgrains. The increase of strain rate weakens the texture during hot compression.

REFERENCES

• 1)   G.B.  Olson: Acta Mater. 61 (2013) 771–781. doi:10.1016/j.actamat.2012.10.045
• 2)   Z.H.  Wang,  B.  Niu,  Q.  Wang,  C.  Dong,  J.C.  Jie,  T.M.  Wang and  T.G.  Nieh: J. Mater. Sci. Technol. 93 (2021) 60–70. doi:10.1016/j.jmst.2021.04.011
• 3)   J.  Tian,  M.B.  Shahzad,  W.  Wang,  L.  Yin,  Z.  Jiang and  K.  Yang: J. Mater. Sci. Technol. 34 (2018) 1671–1675. doi:10.1016/j.jmst.2018.04.020
• 4)   Y.  Li,  W.  Yan,  J.D.  Cotton,  G.J.  Ryan,  Y.  Shen,  W.  Wang,  Y.  Shan and  K.  Yang: Mater. Des. 82 (2015) 56–63. doi:10.1016/j.matdes.2015.05.042
• 5)   J.L.  Tian,  Y.C.  Li,  W.  Wang,  W.  Yan,  Y.Y.  Shan,  Z.H.  Jiang and  K.  Yang: Acta Metall. Sin. 52 (2016) 1517–1526. doi:10.11900/0412.1961.2016.00234
• 6)   J.  Tian,  W.  Wang,  L.  Yin,  W.  Yan,  Y.  Shan and  K.  Yang: Scr. Mater. 121 (2016) 37–41. doi:10.1016/j.scriptamat.2016.04.033
• 7)   J.  Ma,  Q.  Gao,  H.  Zhang,  B.  Wang and  G.  Feng: Mater. Trans. 63 (2022) 1533–1540. doi:10.2320/matertrans.MT-M2022089
• 8)   Y.V.R.K.  Prasad,  H.L.  Gegel,  S.M.  Doraivelu,  J.C.  Malas,  J.T.  Morgan,  K.A.  Lark and  D.R.  Barker: Metall. Trans. A 15 (1984) 1883–1892. doi:10.1007/BF02664902
• 9)   Y.J.  Zhao,  H.C.  Ding,  Y.F.  Cao,  P.L.  Chen,  Z.L.  Hu,  J.R.  Zhang and  L.X.  Li: Metals 11 (2021) 347–360. doi:10.3390/met11020347
• 10)   Q.  Liao,  Y.  Jiang,  Q.  Le,  X.  Chen,  C.  Cheng,  K.  Hu and  D.  Li: J. Mater. Sci. Technol. 61 (2021) 1–15. doi:10.1016/j.jmst.2020.04.064
• 11)   P.  Wan,  H.  Zou,  K.  Wang and  Z.  Zhao: J. Mater. Res. Technol. 13 (2021) 1083–1097. doi:10.1016/j.jmrt.2021.05.019
• 12)   I.  Kartika,  Y.  Li,  H.  Matsumoto and  A.  Chiba: Mater. Trans. 50 (2009) 2277–2284. doi:10.2320/matertrans.M2009103
• 13)   Z.  Jia,  Z.X.  Gao,  J.J.  Ji,  D.X.  Liu,  T.B.  Guo and  Y.T.  Ding: Rare Met. 40 (2021) 2083–2091. doi:10.1007/s12598-020-01474-6
• 14)   L.  Wang,  Z.  Li,  X.  Hu,  B.  Lv,  C.  Chen and  F.  Zhang: J. Mater. Res. Technol. 13 (2021) 618–634. doi:10.1016/j.jmrt.2021.05.006
• 15)   H.  Chen,  Z.  Wang,  F.  Qin,  P.  Jia and  X.  Zhao: J. Wuhan Univ. Technol. Mater. Sci. Ed. 32 (2017) 935–943. doi:10.1007/s11595-017-1693-8
• 16)   M.  Rout,  S.  Biswas,  R.  Ranjan,  S.K.  Pal and  S.B.  Singh: Metall. Mater. Trans. A 49 (2018) 864–880. doi:10.1007/s11661-017-4447-5
• 17)   M.  Eskandari,  M.A.  Mohtadi-Bonab,  A.  Zarei-Hanzaki and  S.M.  Fatemi: J. Mater. Eng. Perform. 27 (2018) 1555–1569. doi:10.1007/s11665-018-3273-2
• 18)   F.  Ren,  F.  Chen,  J.  Chen and  X.  Tang: J. Manuf. Process. 31 (2018) 640–649. doi:10.1016/j.jmapro.2017.12.015
• 19)   Z.  Hu,  K.  Wang and  Y.  Yang: Metall. Microstruct. Anal. 9 (2020) 576–587. doi:10.1007/s13632-020-00662-1
• 20)   H.  Mirzadeh and  A.  Najafizadeh: Mater. Des. 31 (2010) 4577–4583. doi:10.1016/j.matdes.2010.05.052
• 21)   Y.V.R.K.  Prasad: J. Mater. Eng. Perform. 12 (2003) 638–645. doi:10.1361/105994903322692420
• 22)   D.X.  Wen,  Y.C.  Lin,  H.B.  Li,  X.M.  Chen,  J.  Deng and  L.T.  Li: Mater. Sci. Eng. A 591 (2014) 183–192. doi:10.1016/j.msea.2013.09.049
• 23)   Y.B.  Tan,  L.H.  Yang,  C.  Tian,  W.C.  Liu,  R.P.  Liu and  X.Y.  Zhang: Mater. Sci. Eng. A 597 (2014) 171–177. doi:10.1016/j.msea.2013.12.085
• 24)   P.  Zhou,  Q.X.  Ma and  J.B.  Luo: Metals 7 (2017) 50–61. doi:10.3390/met7020050
• 25)   J.D.  Baird: Metall. Rev. 16 (1971) 1–18. doi:10.1179/mtlr.1971.16.1.1
• 26)   S.V.S.  Narayana Murty,  B.  Nageswara Rao and  B.P.  Kashyap: J. Mater. Process. Technol. 166 (2005) 268–278. doi:10.1016/j.jmatprotec.2004.09.089
• 27)   P.  Zhou and  Q.  Ma: Met. Mater. Int. 23 (2017) 359–368. doi:10.1007/s12540-017-6538-7
• 28)   H.W.  Luo,  X.H.  Wang,  Z.B.  Liu and  Z.Y.  Yang: J. Mater. Sci. Technol. 51 (2020) 130–136. doi:10.1016/j.jmst.2020.04.001
• 29)   S.  Morito,  A.H.  Pham,  T.  Hayashi and  T.  Ohba: Mater. Today Proc. 2 (2015) S913–S916. doi:10.1016/j.matpr.2015.07.430
• 30)   P.  Gao,  Z.  Ma,  J.  Gu,  S.  Ni,  T.  Suo,  Y.  Li,  M.  Song,  Y.W.  Mai and  X.Z.  Liao: Sci. China Mater. 65 (2022) 811–819. doi:10.1007/s40843-021-1798-6