Effect of Nitrogen Content on Cold Working Properties of High-strength Mn–Cr–N Steels Made Using Pressurized ESR

Abstract

High-strength Mn–Cr–N steels with high nitrogen content were manufactured using a lab-scale pressurized electro-slag remelting furnace to study the deformability of the steels. Melting experiments were performed under 1.0 MPa pressure N₂ gas in order to have various N contents. Gas porosity and severe macrosegregation were not observed in the remelted ingots. Microstructure observation revealed that nitrides and non-metallic inclusions were small enough not to affect the mechanical properties. After the ESR ingots were heat-treated and forged, the mechanical properties of the steels at a room temperature were measured. The grain sizes were measured in the range from 50 to 300 μm. The results of 0.2% proof stress showed that the steel became stronger with increasing N content according to solid solution hardening mechanism. In addition, with various strain rates, the tensile strain-hardening exponents were determined to be almost the constant values between 0.20 and 0.25. These results suggest that the methods of cold working for conventional 18Mn–18Cr–0.7N steel are applicable to the Mn–Cr–N steels containing over 1.0 mass% nitrogen.

1. Introduction

Nitrogen is well known as a valuable element in steel, improving mechanical properties,^1,2) heat resistance,^3,4) corrosion resistance^5,6) and oxidation resistance.⁷⁾ Many kinds of special steels containing nitrogen have been developed over the past 20 years. Since 1990s, efforts have been made to use nitrogen as an effective austenite stabilizer to address environmental issues and Ni allergies. In particular, Europe has been very active in the research and development of high-nitrogen steels, such as Ni-free high-nitrogen stainless steels.⁸⁾ Some kinds of high-nitrogen steels have already been manufactured commercially.⁹⁾

18Mn18CrN steel is one of these commercial high-nitrogen steels, which is used for the retaining rings of thermal or nuclear large-scale power generator rotors. 18Mn18CrN steel is non-magnetic and has a high stress-corrosion-cracking resistance. Enough strength is ensured due to the addition of nitrogen along with adequate deformation at a room temperature. But the limit of N addition in this steel is approximately 0.7 mass% in either a conventional ladle refining process or an electro-slag remelting (ESR) process.

Recently, it has become practical to add larger amounts of N to molten steel using a pressurized ESR furnace^10,11) or a pressurized induction furnace¹²⁾ with no excessive Cr or Mn addition. Furthermore, these pressurized melting processes could prevent gas porosity formation in the steel ingot during solidification. In Japan, studies of high-nitrogen steel using a pressurized melting or casting process have been reported positively.^13,14,15)

In this paper, we report on the results of manufacturing high-strength Mn–Cr–N steel with a high N content, using a lab-scale pressurized ESR furnace, and we discuss the effects of the N content on the cold working properties of the steel.

2. Experimental Procedures

Figure 1 shows a schematic illustration of the lab-scale pressurized ESR apparatus. The pressurized ESR in this study can cast continuously using a 145 mm diameter water-cooled copper mold mounted on a base plate with a consumable electrode, due to the joule heat generated in the ESR slag. In this study, a cold-start method using solid slags was applied because a hot-start method using premelted slags was inappropriate using the furnace covered with the pressure vessel.

Fig. 1.

Schematic diagram of pressurized ESR apparatus.

ESR experiments were carried out in order to optimize the current conditions before manufacturing Mn–Cr–N steel ingots for evaluation of their mechanical properties. The chemical compositions of the test consumable electrodes are given in Table 1. SUS 304 stainless steel was applied to determine the primary conditions of ESR. The selected ESR slag composition was simple ternary system slag of 60 mass%CaF₂-20 mass%CaO-20 mass%Al₂O₃. N₂ gas was introduced with a pressure of 1.0 MPa into the pressure vessel before the ESR operation with a voltage fixed at 20 V while three current conditions were tested; 2500 A, 2000 A and 1500 A.

Table 1. Chemical composition of test electrode (mass%).

C	Si	Mn	Ni	Cr	Mo	N
0.04	0.34	0.99	8.00	18.30	0.25	0.065

The ingots produced by the ESR were cut longitudinally for macrostructure observation (etching solution: 15 vol%H₂O₂, 43 vol%HCl, and 42 vol% distilled water) and chemical composition analysis. Furthermore, the dendrite arm spacing of the center portions of the ESR ingots were measured to determine the effect of current conditions on the solidification behavior (etching solution 25 vol%HNO₃ and 75 vol%HCl).

Mn–Cr–N steel ingots to study the effect of N contents on mechanical properties were firstly cast using a 50 kg vacuum induction furnace to prepare 100 mm diameter cast electrodes. Table 2 shows the chemical compositions of the cast electrodes. The test steel ingots were thereafter manufactured by a 1.0 MPa nitrogen pressurized ESR process with the electrodes.

Table 2. Chemical composition of test electrodes (mass%).

Steel	C	Si	Mn	Ni	Cr	Mo	N
A	0.06	0.50	23.90	0.51	10.70	0.06	0.370
B	0.09	1.20	17.81	0.21	10.64	3.70	0.292

In this study, N was added to the test steels during the pressurized ESR process. 9 mm diameter ferro chrome-nitride wires were welded to the cast electrode. The chemical compositions of the wires are shown in Table 3. The N and Cr contents of the test steels were controlled by the number of the welded wires. Macrostructures were first observed followed by analysis of distributions of the alloying elements in the Mn–Cr–N steel ESR ingots.

Table 3. Chemical composition of ferro-chrome nitride wire (mass%).

C	Si	Cr	N
0.07	1.42	64.00	7.00

The ESR ingots were heat-teated at 1250°C for 20 h to homogenize the structures and subsequently forged to 100 mm in diameter within a temperature drop from 1250°C to 900°C. Solution annealing was performed at 1100°C or 1150°C for 2 h with air quenching.

The microstructures of the test steels were observed using an optical microscope and a scanning electron microscope (SEM: JSM-6610) equipped with energy dispersive X-ray analysis (EDS: JED-2300) to observe grain structures, precipitates on the grain boundaries and non-metallic inclusions. Crystal texture was observed after heat treatment using electron back scatter diffraction pattern (EBSD) analysis with a Shottky-type field-emission scanning electron microscope (FE-SEM: JSM-7100F) equipped with an EDAX DigiView camera and TSL OIM software, version 6.2.

In this study the following two tests were carried out to determine the cold workability of the steels. One is tensile testing at a room temperature under a constant strain rate of 10⁻²s⁻¹ and the other is compression testing at a room temperature under constant strain rates of 10⁻¹s⁻¹ and 10⁰s⁻¹.

3. Experimental Results

3.1. Investigations of Pressurized ESR Ingots

Figure 2 shows the macrostructure of a longitudinal section of the ESR ingot of 2500 A showing that this ingot is fine without any severe rippling of the surface and gas porosity. The growth of the primary crystal structure was from the ingot surface toward the center above 50 mm from the bottom of the ingot.

Fig. 2.

Macrostructure of longitudinal section of ESR ingot (current: 2500 A).

Figure 3 shows the macrostructure of the ESR ingot of 2000 A. At 40 mm from the ingot bottom, severe necking was merely observed at the ingot surface while no fatal defects were observed at the positions except for the necking. Additionally, the macrostructure of the 2000 A ingot was similar to that of the 2500 A ingot.

Fig. 3.

Macrostructure of longitudinal section of ESR ingot(current: 2000 A).

Figure 4 shows the macrostructure of the ESR ingot of 1500 A with severe rippling at the surface along with entrainment of the ESR slag differing from the other conditions. It was confirmed that the growth direction of the primary crystal structure was perpendicular to the cooled copper base plate implying that the metal pool depth of the 1500 A ESR was extremely shallow.

Fig. 4.

Macrostructure of longitudinal section of ESR ingot (current: 1500 A).

The dendrite structures of longitudinal sections of ESR ingots are shown in Fig. 5. The dendrite structures were clearly observed except in the 1500 A ingot. The secondary dendrite arm spacings were measured in the ranges; from 70 to 170 μm in the 2500 A ingot, from 80 to 140 μm in the 2000 A ingot, and from 90 to 170 μm in the 1500 A ingot. Average N contents of the ingots after the ESR were approximately 0.116 mass%, a 0.05 mass% increase from the original realized in Table1. This implies that solidification mode has not changed that are all ferrite to austenite (FA) described later.

Fig. 5.

Dendrite structures of longitudinal section of ESR ingots.

(a) 2500 A, (b) 2000 A, (c) 1500 A.

The macrostructure of a longitudinal section of the high-strength Mn–Cr–N steel ingot is shown in Fig. 6. The melt rate conditions in the pressurized ESR are summarized in Table 4. The ingot is quite fine without any severe rippling of the surface and nitrogen gas porosity. It was found that the primary crystal structure changed dynamically at a point 50 mm above the bottom of the ingot. From these macrostructure observations, the metal pool depth of the high-strength Mn–Cr–N steel is assumed to be deeper than those of the other steels.

Fig. 6.

Macrostructure of longitudinal section of high-strength Mn–Cr–N steel ESR ingot.

Table 4. Melt rate in the pressurized ESR.

No.	Steel type	Current (A)	Voltage (V)	Melt rate (kg/h)
1	SUS 304	2500	20	42.73
2	SUS 304	2000	20	36.31
3	SUS 304	1500	20	10.36
4	Mn–Cr–N steel	2500	20	38.16

Figures 7 and 8 show Mn and Cr distributions and N distribution in the ESR ingot, respectively. Variation of the alloying elements in the ingot is seen to be low at every sampling position.

Fig. 7.

Mn and Cr distribution in high-strength Mn–Cr–N steel ESR ingot.

Fig. 8.

Nitrogen distribution in high-strength Mn–Cr–N steel ESR ingot.

3.2. Microstructure and Tensile Test Results

The chemical compositions of the test steels are summarized in Table 5. These test steels were prepared using electrodes containing the alloying elements as shown in Table 2. Test steels containing 0.78 mass%, 0.83 mass%, and 1.0 mass% N are designated as A1, A2, and A3, respectively. B1 is the test steel containing 3.0 mass% Mo with a high N content.

Table 5. Chemical compositions of test steels (mass%).

Steel	C	Si	Mn	Ni	Cr	Mo	N
A1	0.06	0.31	18.81	0.50	18.31	0.06	0.777
A2	0.06	0.10	18.93	0.45	18.32	0.05	0.827
A3	0.07	0.07	19.22	0.46	17.33	0.05	1.003
B1	0.09	0.56	13.47	0.22	19.75	3.00	1.205

Figure 9 shows the microstructure of the test steels after solution heat treatment. Austenitic grains were observed in all the specimens. The average grain sizes of the test steels were measured as 142 μm in A1, 125 μm in A2, 149 μm in A3 and 81 μm in B1. This result tells us that the grain size of high-nitrogen steel is considered to be affected by the major alloying elements rather than by N content.

Fig. 9.

Microstructures of test steels after solution heat treatment. (a) A1, (b) A2, (c) A3, (d) B1.

Figure 10 shows secondary electron images of the test steels. A small amount of spinel inclusions (MgO–Al₂O₃) and MnS were observed in all the specimens by the SEM/EDS analysis. Spinel inclusions of the test steels are assumed to be complex nitride-oxide, because these non-metallic inclusions contain some amount of N. The average inclusion size in the test steels was below 5 μm. As a result, it was confirmed that the pressurized ESR in this study has the ability to remove coarse non-metallic inclusions of the consumable electrode.

Fig. 10.

Secondary electron images of test steels. (a) A2, (b) B1.

0.2% proof stress and tensile strength of the test steels are shown in Fig. 11. The values of 0.2% proof stress were over 500 MPa in all the specimens. It is obvious that the 0.2% proof stress and tensile strength increase with an increase in N concentration. Elongation and reduction of area in the test steels are shown in Fig. 12. The elongation values in the test steels are all over 30% which is enough. On the other hand, reduction of area decreases with an increase in N concentration. In addition, ductility anisotropy originated from the forging direction is obvious as shown in Fig. 12.

Fig. 11.

0.2% proof stress and tensile strength in test steels.

Fig. 12.

Elongation and reduction of area in test steels. (a) Longitudinal direction, (b) Transverse direction.

The possibility of texture formation was indicated by the above tensile test results. Figure 13 shows the inverse pole figure maps of the A3 steel measured by EBSD analysis. There was a slight orientation difference between the longitudinal direction (forging direction) and the transverse direction. As shown in Fig. 13(b), EBSD analysis results of the transverse direction indicate low ductility showing that a slip system other than the primary slip system (111) plane in austenitic steel has been activated. It is considered that the test steels forged only to the longitudinal direction still have the trace of the primary crystal structure or forging texture even after the solution heat treatment.

Fig. 13.

Inverse pole figure maps obtain from EBSD analysis in A3 steel. (a) Longitudinal direction, (b) Transverse direction.

True stress-strain curves of the test steels obtained from the tensile tests are shown in Fig. 14. True stress-strain curves of the A3 steel obtained from the compression tests under various strain rates are shown in Fig. 15. It has been reported that the stress-strain curve can be approximated as a sigmoidal shape after the transient strain in 0.39 mass% N containing Fe–18Cr–10Mn alloy.¹⁶⁾ It is well known that this characteristic behavior indicates a deformation-induced martensitic transformation in metastable austenitic alloys. But in this study, as shown in Figs. 14 and 15, the stress-strain curves do not display such behavior.

Fig. 14.

True stress-strain curves of test steels from tensile tests (strain rate: 10⁻² s⁻¹).

Fig. 15.

True stress-strain curves of A3 steel from compression tests at various strain rates.

4. Discussion

4.1. Solidification Behavior in a Pressurized ESR

Channel segregation, called “A” segregation or “freckle”, seldom forms in an ESR ingot, because the temperature gradient in an ESR melt is higher than that in conventional ingot casting. However, the mold size and melt conditions in ESR process dominate the behavior of channel segregation formation similar to conventional ingot casting process due to the formation of coarse solidification structures. As for ESR, minimization of local solidification time has been reported to effectively prevent the formation of channel segregation.¹⁷⁾ For example, Ueda et al.¹⁸⁾ reported the relationship between the local solidification time and melt rate after studying a Ni-based superalloy ESR ingot.

In the macrostructure investigation of this study, channel segregation was not observed in the three SUS 304 ingots. Furthermore, the segregation of the alloying elements classified as positive and negative was not seen clearly. These results indicate that the ESR current conditions in this study maintained the optimum local solidification time and minimized channel segregation formation during the ESR melt.

In order to discuss the local solidification time, the cooling rate was estimated from the secondary dendrite arm spacings of the ESR experimental ingots.

Many empirical equations have been reported for the relation of secondary dendrite arm spacing against cooling rate regarding SUS 304. The equation¹⁹⁾ used in this study, applicable to lower cooling rates, is expressed as follows:   

$$S_2=111.3\cdot V^{-0.45}\left({10}^{-1}<V<{10}^3\right)$$(1)

where S₂ is the secondary dendrite arm spacing (μm), and V is the cooling rate (K/s). N content of the obtained ESR ingots was 0.05 mass% higher than that of the electrode. However, the results of the solidification mode simulation using Thermo-calc indicate that there is little influence on dendrite structure. The average values of the cooling rate in the pressurized ESR were estimated as; 0.87 K/s in the 2500 A melt, 1.02 K/s in the 2000 A melt, and 0.73 K/s in the 1500 A melt. The liquidus-solidus temperature interval calculated by Thermo-calc for the chemical compositions of the ESR experimental ingots was 38.6 K. The local solidification times in the ESR processes were estimated using the assumed cooling rates and the calculated liquidus-solidus temperature intervals.

Figure 16 shows the relationship between the local solidification time and the melt rate in the pressurized ESR. The error bars of the local solidification times are caused by the measurement errors of secondary arm spacings. Within the range of melt rates in the ESR process, it was seen that the local solidification time was almost constant. Therefore, it appears that the high-strength Mn–Cr–N steel ingots were cast under optimum ESR conditions, preventing the freckle formation caused by coarse solidification structure.

Fig. 16.

Relationship between local solidification time and melt rate in pressurized ESR.

4.2. Mechanical Properties of High-strength M–Cr–N Steel

Many reports have explained that the strength of high-nitrogen steel at a room temperature is caused by nitrogen solid solution hardening mechanism. Discussion is made on the effect of solid solution hardening on the high-strength Mn–Cr–N steels prepared. First, because the test steels have various grain sizes, the grain size effect on the strength has been determined at a room temperature.

The correlation between the Hall-Petch parameter and the N concentration has been reported by Tsuchiyama et al.²⁰⁾ It is well known that the Hall-Petch parameter in stainless steels containing more than 0.4 mass% N is almost constant in the range of 0.80 to 0.90 MPa·m^1/2. The grain size effect on the 0.2% proof stress was evaluated using a certain Hall-Petch parameter, taken as 0.85 MPa·m^1/2. The Hall-Petch equation is expressed as follows:   

$${\sigma }_{0.2}={\sigma }_0+k\cdot d^{-1/2}$$(2)

where σ_0.2 is the 0.2% proof stress (MPa), σ₀ is the material constant (MPa), k is the Hall-Petch parameter (MPa·m^1/2), d is the grain diameter (m). The value of k·d^‒1/2 indicates the strength increment due to grain refinement when the material constants σ₀ of test steels are the same. Figure 17 shows the effect of grain size on the 0.2% proof stress in the test steels. The results show that the values of the 0.2% proof stress are mainly dominated by solid solution hardening based on the N concentration.

Fig. 17.

Effect of grain size on 0.2% proof stress in test steels.

The relationship between the 0.2% proof stress values and N contents is shown in Fig. 18. The values reported for a Fe–Cr–Mn–N alloy²¹⁾ are presented together with our work in Fig. 18. In the high-nitrogen region, it was reported that the 0.2% proof stress is proportional to the atomic fraction of N to the power of 2/3. Our results were confirmed to be almost the same as the previous results proving that our values of 0.2% proof stress can be explained by the theory of nitrogen solid solution hardening.

Fig. 18.

Relationship between 0.2% proof stress and nitrogen content.

The influence of N content on cold working properties is discussed. The results of the tensile and compression tests were analyzed by the same manner described below, because the shape changes of the compression tests indicated that the friction coefficient was zero.

For a single-axis deformation, the true stress σ and true strain ε are calculated using the classical equation²²⁾ as follows:   

$$\sigma =\left(1+e\right){\sigma }_n,\epsilon =ln\left(1+e\right)$$(3)

where σ_n is the nominal stress, and e is the nominal strain. It should be noted that this equation cannot be applied in the instable plastic deformation region corresponding to the necking formation. The strain hardening exponent was estimated using the experimental values in the range of the stable plastic deformation region. Furthermore, the logarithm values of the true stress and true strain are expressed as a linear relation of a power law model as follows:   

$$\sigma =C{\epsilon }^n$$(4)

where C is the constant, and n is the strain hardening exponent.

Figure. 19 shows the relationship between the strain hardening exponent and the N content (strain rate: 10⁻²s⁻¹). As shown in Fig. 19, the strain-hardening exponent values are mostly the same as each other and determined to be between 0.20 and 0.25.

Fig. 19.

Relationship between strain-hardening exponent and nitrogen content (strain rate: 10⁻² s⁻¹).

Many reports have discussed the work hardening properties of high-nitrogen steel, applying the behavior of stacking fault energy²³⁾ or the formation of N–Cr or N–Mo atom pairs.^24,25) However, the constant values of the strain-hardening exponents suggest that the extreme dislocation multiplication behavior, the changing of the stacking fault energy and the shift of the deformation mechanism do not occur in the range of N concentration of the test steels. The results similar to this study were reported by Nam et al. in a Fe–18Cr–14Mn–4Ni–3Mo alloy.²⁶⁾ The strain-hardening exponents determined by a modified Ludwik model (strain rate: 5×10⁻²s⁻¹), in the range from 0.51 mass% to 0.88 mass% N content, are almost constant values between 0.29 and 0.31.

Figure 20 shows the relationship between the strain-hardening exponent and strain rate in the A3 steel. Nam et al.²⁶⁾ reported the increase in the strain-hardening exponent for a low strain rate (5×10⁻⁵s⁻¹). However, the effect of strain rate is not clearly seen in the present work. The strain-hardening exponents of the test steels are between 0.23 and 0.30.

Fig. 20.

Relationship between strain-hardening exponent and strain rate in A3 steel.

As a result, it is suggested that the method of cold working for the conventional 18Mn–18Cr–0.7N steel is applicable to the high-strength Mn–Cr–N steels containing N over 1.0 mass% under strain rates between 10⁰ and 10⁻²s⁻¹.

5. Conclusions

In this study, high-strength Mn–Cr–N steels with high N contents were firstly cast using a lab-scale pressurized ESR furnace. The effect of N content on the cold working properties was determined. The obtained results can be summarized as follows.

(1) The macrostructures of the high-strength Mn–Cr–N steel ingots showed that the ingots were sound without channel segregation and severe segregation of alloying elements.

(2) The microstructures of the test steels after solution heat treatment showed that internal quality was fine enough without coarse inclusions and nitride precipitation and that the grain sizes were from 50 to 300 μm.

(3) The 0.2% proof stress values of the high-strength Mn–Cr–N steels was proportional to the atomic fraction of N to the power of 2/3. This proved that hardening behavior obeyed nitrogen solid solution hardening mechanism.

(4) The strain-hardening exponents of the high-strength Mn–Cr–N steel were determined to be almost constant between 0.20 and 0.25.

(5) Under strain rates from 10⁰ to 10⁻²s⁻¹, the method of cold working for the conventional 18Mn–18Cr–0.7N steel is applicable to the high-strength Mn–Cr–N steels containing N over 1.0 mass%.

References

• 1)   M. O.  Speidel: High Nitrogen Steels, Stahleisen, Düsseldorf, (1990), 128.
• 2)   P. G.  Uggowizer and  M.  Harzermoser: High Nitrogen Steels, Institute of Metals, London, (1988), 174.
• 3)   K.  Hara,  H.  Aoki,  F.  Masuyama and  T.  Endo: ISIJ Int., 37 (1996), 181.
• 4)   A.  Goecmen,  R.  Steins,  C.  Solenthaler,  P. J.  Uggowitzer and  M. O.  Speidel: ISIJ Int., 36 (1996), 768.
• 5)   M.  Sagara,  Y.  Katada and  T.  Kodama: CAMP-ISIJ, 15 (2002), 442.
• 6)   S.  Azuma,  H.  Miyuki and  T.  Kudo: ISIJ Int., 36 (1996), 793.
• 7)   F.  Masuyama,  K.  Hiromatsu and  Y.  Hasegawa: ISIJ Int., 36 (1996), 825.
• 8)   D.  Kuroda: Materia Jpn., 43 (2004), 643.
• 9)   G.  Stein and  J.  Menzel: Int. J. Mater. Prod. Tech., 10 (1995), 290.
• 10)   G.  Stein and  J.  Menzel: Int. J. Mater. Prod. Tech., 10 (1995), 478.
• 11)   Y.  Katada,  M.  Sagara,  Y.  Kobayashi and  T.  Kodama: CAMP-ISIJ, 15 (2002), 438.
• 12)   T.  Sugiyama and  T.  Shibata: Denki Seiko, 76 (2005), 5.
• 13)   Y.  Katada: Materia Jpn., 45 (2006), 444.
• 14)   S.  Hamano,  T.  Shimizu and  T.  Noda: Denki Seiko, 77 (2006), 107.
• 15)   S.  Narita,  S.  Hamano and  T.  Shimizu: Denki Seiko, 77 (2006), 171.
• 16)   S. J.  Kim,  T. H.  Lee and  C. S.  Oh: Steel Res. Int., 80 (2009), 467.
• 17)   A.  Mitchell: Mater. Sci. Eng., A413–414 (2005), 10.
• 18)   S.  Ueda,  M.  Funazaki,  J.  Taira and  H.  Tamada: Superalloys 718,625,706 and Various Derivatives, TMS, Warrendale, PA, (2001), 71.
• 19)   H.  Esaka,  S.  Mizoguchi,  H.  Kajioka and  H.  Chida: CAMP-ISIJ, 1 (1988), 268.
• 20)   T.  Tsuchiyama,  S.  Takaki and  K.  Kataoka: CAMP-ISIJ, 18 (2005), 1532.
• 21)   T.  Tsuchiyama,  H.  Ito,  K.  Kataoka and  S.  Takaki: Metall. Mater.Trans., 34A (2003), 2591.
• 22)   J. H.  Hollomon: AIME Met. Tech., (1945), 1879.
• 23)   M.  Ojima,  Y.  Adachi,  Y.  Tomota,  Y.  Katada,  Y.  Kaneko,  K.  Kuroda and  H.  Saka: Steel Res. Int., 80 (2009), 477.
• 24)   M.  Murayama,  K.  Hono,  H.  Hirukawa,  T.  Ohmura and  S.  Matsuoka: Scr. Mater., 41 (1999), 467.
• 25)   M.  Yoshitake,  T.  Tsuchiyama and  S.  Takaki: Tetsu-to-Hagané, 98 (2012), 223.
• 26)   S. M.  Nam and  Y. S.  Kim: Materials Science and Technology, MS&T, Michigan, (2007), 4086.