Rapid Oxynitriding of Ti–6Al–4V Alloy by Induction Heating in Air

Abstract

In this study, a rapid oxynitriding technique based on induction-heating was developed for the titanium alloy Ti–6Al–4V, and its effects on surface characteristics of the alloy were investigated. The surface microstructures of the alloy were characterized by employing a scanning electron microscope, electron back-scatter diffraction, X-ray diffraction and nano-indentation tests. Induction-heating in air for 60 s caused the formation of a hardened layer on the Ti–6Al–4V alloy surface, which can be attributed to the diffusion of oxygen and nitrogen atoms. When water quenching was performed after heating, transformation to an acicular martensitic α′ phase occurred, resulting in increased in substrate hardness. However, cracks were initiated at the surface of the oxynitrided alloy. The cracks were eliminated through fine particle bombarding treatment, whereas the hardened layer formed by the oxynitriding treatment remained. The obtained results thus indicate that the oxynitriding technique developed in this study can enhance the wear resistance and tensile strength of the Ti–6Al–4V alloy within a short period of time.

 

This Paper was Originally Published in Japanese in J. Soc. Mater. Sci., Jpn. 69 (2020) 605–611.

Fig. 1 Results of EBSD analysis (IPF, phase and IQ maps) of W, Q and A materials.

1. Introduction

Ti–6Al–4V is a representative titanium alloy with α + β phase. This alloy possesses high specific strength, superior corrosion resistance, and biocompatibility. Thus, it is utilized in the aerospace and medical fields. However, the material exhibits poor wear resistance. Hence, to improve wear resistance, surface hardening treatments, such as nitriding, oxidation, and oxynitriding, were performed, and their effects were examined.^1–9) However, these treatments require long treatment times that generally last for several hours. Titanium alloys can be applied in a wider range of fields if improvements in productivity can be achieved via reducing its long treatment time.

In this study, we focus on induction-heating (IH) to form hardened layers on titanium alloy surfaces within a shorter span of time when compared to that of conventional methods. Specifically, IH can introduce high energies in a short time, and thus, realize rapid heating to high temperatures. The authors previously revealed that nitrided layers were formed at the surfaces of commercially pure titanium and Ti–6Al–4V alloy for 3 min via IH in the chamber replaced with a nitrogen atmosphere.^10–12) This implies that IH for titanium at high temperatures can rapidly diffuse elements in ambient atmosphere. Therefore, in this study, a rapid oxynitriding technique is developed for titanium alloys via IH in air without chambers and diffusion of oxygen and nitrogen atoms.

It was reported in extant studies that quenching of titanium alloys after IH results in the formation of α′ martensite phase and increases the hardness and tensile strength.^13,14) Based on these results, there is a possibility that quenching after IH in air forms hardened layers at the surface and α′ phase inside the material. This results in simultaneous improvement of the wear resistance and tensile strength of titanium alloys.

Based on these backgrounds, a rapid oxynitriding process for titanium alloy is developed via IH at high temperatures in air. This leads to rapid diffusion of oxygen and nitrogen. Furthermore, quenching after IH was conducted to transform the microstructure and increase the internal hardness. The possibility of improving the wear resistance and tensile strength via the proposed treatment is discussed based on microstructure observations and hardness measurements of the treated alloy.

2. Experimental Procedures

2.1 Specimen preparation

In this study, rolled bars of Ti–6Al–4V alloy were used. The chemical composition of the alloy is shown in Table 1. The bars were machined into specimens with a diameter of 10 mm and length of 40 mm. Furthermore, the specimens were polished using emery papers and a colloidal silica suspension to obtain a mirror finish. This polished specimen is referred to as W material in this study. Finally, as per the conditions listed in Table 2, IH was performed on the W material. The treatment temperatures and cooling methods were varied, and their effects on the microstructures of the alloy were investigated. The heating time was set based on a previous study,¹³⁾ which reported strengthening of the Ti–6Al–4V alloy via heat treatment with a short time. The treatment temperature was set above and below the β transus point of Ti–6Al–4V alloy (1268 K). Hereafter, the water-quenched and air-cooled specimens are referred to as the Q and A materials, respectively. Each specimen is abbreviated by denoting the treatment temperature before Q and A, i.e., 1173A material.

Table 1 Chemical composition of Ti–6Al–4V alloy, mass%.

Table 2 Conditions of induction heating.

2.2 Characterization of the surface-modified layer

Button-shaped specimens with a length of 5 mm were machined from the center of the round-bar-shaped W, Q, and A materials, and the following analyses were performed.

The microstructures were analyzed via a scanning electron microscope (SEM) equipped with electron back-scattered diffraction (EBSD) and by obtaining inverse pole figure (IPF), image quality (IQ) and phase maps.

The compounds formed on the surfaces were identified via X-ray diffraction. The X-ray diffraction was conducted with Cu Kα radiation, diffraction angle, 2θ, in the range of 20° to 60°, scan step of 0.01°, and scan speed of 0.25°/min.

The hardness distribution was measured via nano-indentation tests. A Berkovich indenter with a ridge angle of 115° was used with a testing load of 45 mN. This measurement was conducted in the radial direction of the button-shaped specimen, and the distance from the surface ranged from 10 to 510 µm. The resulting hardness value was defined as the average of the three measurements at each depth. In this study, hardness is denoted by HT115, as defined by eq. (1).   

\begin{equation} \text{HT115} = 160.07 \times F/L^{2} \end{equation} (1)

Where, F denotes the testing load and L denotes the height of the triangle indentation. The hardness of the material is determined by measuring the size of the triangle indentations and dividing the surface area of the indentation by the testing load. Thus, this calculation corresponds Vickers hardness.

3. Results and Discussion

3.1 Effect of IH in air on the microstructure of titanium alloys

Figure 1 shows the results of the EBSD analysis conducted on the W, Q, and A materials. Specifically, the W material was mainly composed of equiaxed α phase and exhibited the same microstructure on the surface and inside. In the 1323Q and 1423Q materials, the microstructure near the surface exhibited coarsened α phase, which became thicker as the treatment temperature was increased. Oxygen and nitrogen, which diffuse into titanium during this treatment, are α-stabilized elements. Therefore, α-case layers are formed due to their diffusion.¹⁵⁾ Given that the thickness of the α-case layer is almost consistent with the depth at which oxygen and nitrogen diffuse, it is presumed that these elements diffused more deeply as the treatment temperature increased. Furthermore, the phase maps of these specimens revealed that β phase existed below the α-case layer. This is because V, which is a β-stabilized element and uniformly exists in the alloy, was concentrated below the α-case layer due to the diffusion of oxygen and nitrogen. In the internal cross-section of the 1323Q and 1423Q materials, an acicular α′ phase was formed. This is due to the martensitic transformation owing to water-quenching. In a previous study,¹⁶⁾ we reported on the possibility that fine β phase exists in the dark area of the phase map. Thus, it is inferred that the β phase was generated in the fine acicular α′ phase. Conversely, the α-case layer and uniform acicular α′ phase were not formed in the 1173Q material. This is because the diffusion of oxygen and nitrogen and uniform martensitic transformation did not occur at the low treatment temperature.

Fig. 1

Results of EBSD analysis (IPF, phase and IQ maps) of W, Q and A materials.

The 1173A material exhibited the same microstructure as the W material. This implies that the heat treatment did not affect the microstructure. In the 1323A and 1423A materials, coarsened α phase was observed at the surface. This is because equal amounts of oxygen and nitrogen diffused irrespective of the cooling method when the treatment temperature was same. The grain sizes of the α phase near the surfaces of the 1323Q, 1323A, 1423Q, and 1423A materials were measured via the intercept method, and their grain sizes were calculated as 5.0, 8.0, 9.7, and 24.3 µm, respectively. This suggests that grain coarsening occurred in the A materials as opposed to that in the Q materials when the treatment temperature was same. This is because the specimens were maintained at high temperatures for a long time due to the lower cooling rate of air-cooling. In contrast, the microstructure inside the A materials exhibited coarsened α phase and small amounts of acicular α′ phase, which was different from that of the Q materials. In Fig. 1, the interfaces between the α phase near the surface and substrate are indicated by white dashed lines. Given the lower cooling rate of air-cooling when compared with that of water-quenching, the transformation to α′ phase did not occur completely, and most regions exhibited α phase and partly an acicular α′ phase.

Figure 2 presents the X-ray diffraction profiles of the W, Q, and A materials. The diffraction peaks of TiO₂ were detected from the profiles of the 1173Q and 1173A materials. However, significant peaks of oxygen or nitrogen compounds were not detected in specimens treated at higher temperatures although a weak peak of TiN was detected in the 1423Q material. Hence, it is inferred that compound layers were formed only at the surfaces of the specimens treated at 1173 K, irrespective of the cooling rate. This is probably because the diffusion rate of oxygen and nitrogen was faster than the rate required for forming compounds with Ti. Therefore, oxygen and nitrogen diffused inward without the formation of compounds. A detailed analysis of the Ti α peaks revealed that the peaks from the induction-heated specimens shifted to lower angles when compared with those from the W material. This shows the diffusion of oxygen and nitrogen atoms in titanium. During this treatment, oxidization occurred preferentially because TiO₂ was mainly formed. This is consistent with the result of a previous study¹⁷⁾ wherein the diffusion rate of oxygen in titanium is much higher than that of nitrogen at 973–1273 K.

Fig. 2

X-ray diffraction profiles of W, Q and A materials.

Figure 3 shows the hardness distributions measured along the longitudinal sections of each specimen. The broken line in this figure shows the hardness of the W material (334.8 ± 7.2 HT115, n = 63, mean ± S.D.). Specifically, the Q and A materials exhibited high hardness values near the surface. This is attributed to the diffusion of oxygen and nitrogen during IH. The thickness of the hardened layer and surface hardness value increased as the treatment temperature increased. This is due to the accelerated diffusion of oxygen and nitrogen. Moreover, the hardness at the outermost surface was approximately the same for each treatment temperature irrespective of the cooling method. This is because the hardness at the outermost surface is dominated by the extent of elemental diffusion during IH. Below the surface, the hardness of both materials decreased rapidly within a distance of several tens of micrometers from the surface. This length corresponds to the thickness of the α-case layer shown in Fig. 1. This decrease in hardness is due to the decrease in oxygen and nitrogen concentrations. At a deeper hardened layer, there was a difference in the hardness between the Q and A materials, and this difference became more significant as the treatment temperature increased. This is due to the microstructural transformation of the Q material to α′ phase via quenching. The substrate where the hardness value was almost constant also increased in the Q and A materials when compared with that of the W material. This was due to the formation of the acicular α′ phase. This tendency is also observed at the center of the button-shaped specimens.

Fig. 3

Hardness distributions measured from surface for W, Q and A materials.

Takahashi et al.⁶⁾ revealed that a hardened layer with a thickness of approximately 20 µm was formed by heating Ti–6Al–4V alloy at 973 K for 18 ks in air. Although the treatment temperature in this study is higher than in the study by Takahashi et al., a hardened layer with a maximum thickness of approximately 200 µm was obtained for 60 s. Therefore, IH in air is an effective technique as rapid surface hardening treatment of titanium alloys.

Based on the aforementioned results, the hardened layers can be rapidly formed at the surface of the titanium alloys via IH in air. Furthermore, the internal hardness increases via quenching after IH due to the formation of the α′ phase. However, cracks were observed at the surfaces of the treated specimens. Figure 4 shows SEM micrographs of the observed cracks. The crack lengths of the Q and A materials increased as the treatment temperature increased. This crack initiation is attributed to the difference in strain between the surface and inside of the material during cooling. Thus, higher difference in strain led to the generation of longer cracks. Hence, it is inferred that crack length increases as the treatment temperature increased. We expected the inhibition of crack initiation in the A material when compared to that in the Q material because the cooling rate of air-cooling is lower than that of water-quenching, and thus, the difference in strain between the surface and inside is mitigated. However, the crack length of the A material was almost the same as that of the Q material. Therefore, it can be concluded that crack initiation is not inhibited by the type of cooling method.

Fig. 4

SEM micrographs of the surfaces of Q and A materials.

Thus, to utilize the rapid oxynitriding technique developed in this study, it is necessary to inhibit the initiation of cracks or eliminate the initiated cracks, which lead to the reduction in fatigue strength.

3.2 Elimination of the cracks via fine particle bombarding

To eliminate the cracks initiated at the surface treated with rapid oxynitriding, fine particle bombarding (FPB) was introduced. In this treatment, fine particles with a diameter of less than 200 µm are accelerated using compressed air and shot to the surface of the material. In our previous studies,^18–21) we revealed that FPB could eliminate compound layers formed at the outermost surfaces and generate compressive residual stresses after elemental diffusion treatments such as nitriding. Thus, the initiated cracks can be rendered as harmless via FPB treatment by eliminating the surface layers and introducing compressive residual stresses.

Based on previous studies,^18–21) FPB was performed using the following three steps. First, SiC particles were used to remove surface layers; second, SKH51 particles were used to introduce compressive residual stresses; and finally, SiO₂ particles were used to reduce surface roughness. In each step, the injection pressure was set to 0.4 MPa, and the treatment time was 10 s. The mesh sizes corresponded to 400 for SiC and SiO₂ particles and 300 for SKH51 particles. Hereafter, the specimens treated with FPB after oxynitriding are denoted by the abbreviation of oxynitrided materials (F), i.e., 1173AF material.

Figure 5 shows the results of the EBSD analysis of the W, QF, and AF materials. In the figure, the dashed lines show the surfaces of the specimens. The effect of FPB is within several micrometers from the surface.²²⁾ Therefore, the internal microstructures of the specimens treated with FPB were the same as those of the Q and A materials, as shown in Fig. 1. In contrast, dark regions existed near the surfaces because the grains were refined by FPB, and the results of the EBSD analysis were not adequately obtained. However, the α-case layers, where oxygen and nitrogen were diffused, remained after FPB.

Fig. 5

Results of EBSD analysis (IPF, phase and IQ maps) of W, QF and AF materials.

Figure 6 demonstrates the X-ray diffraction profiles obtained from each specimen. The X-ray diffraction profiles of the 1173QF and 1173AF materials did not exhibit peaks of TiO₂, which were observed in the 1173Q and 1173A materials (Fig. 2). This suggests that the compound layers were eliminated via FPB. The peaks of the specimens treated with FPB were broadened and compared with those without FPB. This is due to grain refinement near the surfaces and generation of residual stress. Furthermore, part of the specimen exhibited peaks of Fe₂O₃. This is because iron oxides contained in SKH51 particles used during FPB were transferred to the treated surfaces.²³⁾

Fig. 6

X-ray diffraction profiles of W, QF and AF materials.

Figure 7 shows the hardness distributions measured along the longitudinal sections of the FPB-treated specimens. The outermost hardness of the QF and AF materials decreased when compared with those of the Q and A materials (Fig. 3). This is attributed to the elimination of the hardened layers via FPB. Based on the comparison of the hardness distributions of the specimens before and after FPB, the thickness of the eliminated layer is estimated to be approximately 20–30 µm, as shown in Figs. 3 and 7. However, the hardened layers remained in the specimens that were oxynitrided at 1323 and 1423 K. This indicates a possibility that the wear resistance is improved by the formation of the hardened layer in the case when FPB is performed. The internal hardness of each specimen was same before and after FPB, and the QF materials exhibited higher hardness than the W and AF materials due to the formation of the α′ phase.

Fig. 7

Hardness distributions measured from surface for W, QF and AF materials.

Figure 8 presents the SEM micrographs of the cracks observed in the specimens treated with FPB. After FPB, it was qualitatively inferred that the length of the cracks decreased and crack closure occurred when compared with the specimens before FPB (Fig. 4). This is because the surface layers, including the cracks, were removed and compressive residual stresses were generated. To quantitatively investigate the effect of FPB on the crack length, the lengths were measured via SEM. The measured area was 1 mm at the circumference of the specimens, and the length of the maximum crack was measured. Figure 9 shows the results. In all the oxynitrided specimens, maximum crack length was decreased via FPB. The reduction rate of the maximum crack length varied among the specimens. However, the reason for this was not apparent. The examination of the FPB conditions, which can completely eliminate cracks, is also a topic of a future study.

Fig. 8

SEM micrographs of the surfaces of QF and AF materials.

Fig. 9

Maximum crack depth formed at the surfaces of the specimens before and after FPB treatment.

The oxynitriding proposed in this study and a post-treatment with FPB can increase surface and internal hardness values of titanium alloys and decrease the length of the cracks, which reduce fatigue strength. This indicates that the proposed treatment is effective in rapidly improving the wear resistance and tensile strength of the alloy. In future studies, detailed effects will be investigated via quantitative analysis of oxygen and nitrogen concentration using electron probe micro analyzer, wear tests and tensile tests.

4. Conclusions

A novel oxynitriding technique for Ti–6Al–4V alloys was developed utilizing IH in air to rapidly form a hardened layer at the surface by diffusing oxygen and nitrogen and simultaneously increasing the internal hardness via microstructural transformation. The main conclusions are as follows.

1. (1)    A hardened layer was formed at the surface of the Ti–6Al–4V alloys by performing IH above 1173 K in air for 60 s. In this treatment, the surface hardness value and thickness of the hardened layer increased as the treatment temperature increased.
2. (2)    In the aforementioned treatment, the internal hardness of the treated alloy was also increased via quenching after IH. This is due to the microstructural transformation to the α′ martensite phase.
3. (3)    Cracks were initiated at the surface of the water-quenched alloy. The length of the crack increased as the treatment temperature increased. Furthermore, crack initiation was not inhibited in the case of air-cooling, which exhibited a decreased cooling rate.
4. (4)    FPB treatment after IH in air led to a decrease in crack length at the surface. Furthermore, crack closure was observed in the remaining parts of the hardened layer. This was potentially due to the elimination of the layer containing cracks via collision with the shot particles.

Acknowledgments

This study has been supported by the JSPS KAKENHI (Grant Number 19H02027). We also thank Ms. Maho Morikawa (Kyoto Institute of Technology) for the help with the experiments.

REFERENCES

• 1)   T.  Yashiki and  F.  Kamikubo: J. JILM 50 (2000) 577–583.
• 2)   J.R.  Cuthill,  W.D.  Hayesand and  R.E.  Seebold: J. Res. Natl. Bur. Stand., A Phys. Chem. 64A (1960) 119–125.
• 3)   M.  Rahman,  I.  Reid,  P.  Duggan,  D.P.  Dowling,  G.  Hughes and  M.S.J.  Hashmi: Surf. Coat. Technol. 201 (2007) 4865–4872.
• 4)   A.F.  Yetim,  F.  Yildiz,  Y.  Vangolu,  A.  Alsaran and  A.  Celik: Wear 267 (2009) 2179–2185.
• 5)   A.R.  Ebrahimi,  F.  Zarei and  R.A.  Khosroshahi: Surf. Coat. Technol. 203 (2008) 199–203.
• 6)   K.  Takahashi and  E.  Sato: Mater. Trans. 51 (2010) 694–698.
• 7)   S.  Banfield,  J.C.A.  Wilson,  G.  Cassar,  A.  Leyland,  A.  Matthews and  J.  Housden: Surf. Coat. Technol. 206 (2011) 1955–1962.
• 8)   T.  Wierzchoń,  E.  Czamowska,  J.  Grzonka,  A.  Sowińska,  M.  Tarnowski,  J.  Kamiński,  K.  Kulikowski,  T.  Borowski and  K.J.  Kurzydłowski: Appl. Surf. Sci. 334 (2015) 74–79.
• 9)   O.I.  Yaskiv,  I.M.  Pohrelyuk,  V.M.  Fedirko,  D.B.  Lee and  O.V.  Tkachuk: Thin Solid Films 519 (2011) 6508–6514.
• 10)   S.  Ota,  S.  Kikuchi,  H.  Akebono,  M.  Omiya,  J.  Komotori,  K.  Fukazawa,  Y.  Misaka and  K.  Kawasaki: J. Jpn. Soc. Abras. Technol. 59 (2015) 643–648.
• 11)   S.  Takesue,  S.  Kikuchi,  H.  Akebono,  J.  Komotori,  K.  Fukazawa and  Y.  Misaka: Mater. Trans. 58 (2017) 1155–1160.
• 12)   S.  Takesue,  S.  Kikuchi,  H.  Akebono,  Y.  Misaka and  J.  Komotori: Surf. Coat. Technol. 359 (2019) 476–484.
• 13)   T.  Morita,  W.  Niwayama,  K.  Kawasaki and  Y.  Misaka: Trans. Jpn. Soc. Mech. Eng. A 64 (1998) 2115–2120.
• 14)   T.  Morita,  Y.  Misaka,  K.  Kawasaki and  T.  Iizuka: J. Jpn. Inst. Metals 68 (2004) 862–867.
• 15)   K.  Farokhzadeh and  A.  Edrisy: Mater. Sci. Eng. A 620 (2015) 435–444.
• 16)   T.  Morita,  C.  Tsuda,  H.  Sakai and  N.  Higuchi: Mater. Trans. 58 (2017) 1397–1403.
• 17)   M.  Yanagisawa,  M.  Nakamura and  T.  Hibino: J. JILM 1955 (1955) 154–159.
• 18)   T.  Morita,  T.  Ohtomo,  C.  Kagaya,  S.  Tanaka and  N.  Tsuji: J. Soc. Mater. Sci. Jpn. 56 (2007) 872–879.
• 19)   T.  Morita,  T.  Ohtomo,  C.  Kagaya,  S.  Tanaka and  N.  Tsuji: J. Soc. Mater. Sci. Jpn. 57 (2008) 838–845.
• 20)   T.  Morita,  N.  Uehigashi and  C.  Kagaya: Mater. Trans. 54 (2013) 1719–1724.
• 21)   T.  Morita,  K.  Asakura and  C.  Kagaya: Mater. Sci. Eng. A 618 (2014) 438–446.
• 22)   S.  Takagi,  M.  Kumagai,  Y.  Ito,  S.  Konuma and  E.  Shimodaira: Tetsu-to-Hagané 92 (2006) 318–326.
• 23)   Y.  Kameyama and  J.  Komotori: J. Mater. Process. Technol. 209 (2009) 6146–6155.