Crevice Corrosion Resistance and Structure of Passive Film on Fe–Mn–Si–Cr–Ni Steel

Abstract

The crevice corrosion resistance of an Fe–Mn–Si–Cr–Ni (15 Mn) steel as a shape memory alloy was estimated by laser microscopy and statistical calculation of the corrosion depth, and the structure of the passive film formed on the steel was examined by EELS (Electron Energy Loss Spectroscopy) analysis using TEM (transmission electron microscopy). The crevice corrosion results were analyzed using a Gumbel distribution, and the mode (λ) and distribution parameter (α) of the maximum corrosion depth of the 15 Mn steel were found to be much smaller than those of 430 stainless steel (SUS: 16Cr – Fe). This shows that 15 Mn steel has a higher crevice corrosion resistance than 430 SUS steel. In AES and XPS analysis, the passive film of 15 Mn steel was shown to contain Fe, Mn, Cr, Si and Ni. From TEM-EELS, the passive film was found to consist of 2 layers. In the passive film, Cr and Si are thought to be effective early, with Ni preventing penetration of crevice corrosion. It was found that 15 Mn steel could maintain a passive film which was composed of effective elements in a crevice corrosion environment.

1. Introduction

In a shape memory alloy,^1,2,3) iron-based alloys have garnered attentions because of reduced material costs. Although Fe–Mn–Si based shape memory alloys have been reported,^4,5) there are few papers covering their corrosion resistance.^6,7) Recently, an attempt to apply an iron-based shape memory alloy in the vibration dampers in tower buildings,⁸⁾ where materials are exposed to atmospheric corrosion, was reported. In a previous paper,⁹⁾ the atmospheric corrosion resistance of this alloy was investigated in the wet and dry cyclic test with a chloride solution. In general, like stainless steels, this type of alloy makes the passive film in many environments. Thus, it is important to examine the crevice corrosion resistance because the passive film is sensitive to this type of corrosion. In this study, corrosion depth was measured quantitatively using laser microscopy and estimated statistically by Gumbel distribution analysis^10,11,12) after crevice corrosion test. In the crevice corrosion, the deepest depth determines the lifetime of the material, because the first leak originates from the deepest depth.

Although the analyses of the corrosion product and the passive film of steels have been conducted, it is the problem that in-situ observation cannot be conducted for them. Recently, the Focused Ion Beam (FIB) method has been proposed to solve this problem.^{13,14,15,16,17)} By using this FIB method for samples, TEM – EELS¹⁸⁾ was additionally applied to analyze the chemical state of the passive film of the steel. Thus, in –situ observation can be applied for the passive film of steel in this study. As the passive film was very thin (~nm), nano structure of the film was investigated by using TEM. Especially, the upper and lower passive films were separately examined, and the chemical state of each element was investigated by EELS. EIS was employed to investigate the crevice corrosion behavior of samples.¹⁹⁾ Finally, the relationship between the crevice corrosion behavior and the structure of the passive film on the 15 Mn steel was discussed.

2. Experimental Procedure

2.1. Crevice Corrosion Test and Samples

Crevice corrosion tests were conducted on the samples of 40×40 mm in size in 15 mass% NaCl solution at 60°C using multi-crevice type jigs for 4 weeks. The sample material (15 Mn steel) was an Fe - 15% Mn - 10% Cr - 4% Si - 8% Ni (mass%) shape memory alloy. As a basis for comparison, the 430 stainless steel (SUS 430; Fe-16%Cr) was used in this study.

2.2. Gumbel Distribution Analysis

In order to identify the crevice corrosion of steels, statistical analysis was introduced. To evaluate the corrosion depth of steels, Gumbel distribution analysis was applied as in previous papers.^10,11,12) As the deepest depth of the crevice corrosion is thought to determine the life of a steel, Gumbel distribution analysis was applied in this study after removing the rust from 15 Mn steel and SUS 430 steel.

The Gumbel distribution has the following form:   

$${\text{F}}_{\text{I}}\left(\text{x}\right)=\text{exp }\left(-\text{exp}\left(-\left(\text{x}-\lambda \right)/\alpha \right)\right)$$(1)

where F_I (x) is the probability function of x and the constants λ and α are called the scale and location parameters, respectively. If a reduced variate defined by   

$$\text{y}=\left(\text{x}-\lambda \right)/\alpha$$(2)

is introduced, Eq. (1) can be expressed as   

$${\text{F}}_{\mathrm{I}}\left(\text{y}\right)=\text{exp}\left(-\text{exp}\left(-\text{y}\right)\right)$$(3)

Then the explicit expression of y is given by   

$$\text{y}=-\text{ln }\left(-\text{ln}\left({\text{F}}_{\text{I}}\left(\text{y}\right)\right)\right)$$(4)

The Gumbel probability diagram may be constructed with values scaled on the vertical (y) axis and the maximum corrosion depth on the horizontal (x) axis.

In the actual experiment, the surface area is divided into 8 equal units on each side of sample (40×40 mm). The maximum depths are measured by a laser microscope for 6 unit areas in a sample, and summarized using the Gumbel probability figure. In order to determine λ and α, MVLUE²⁰⁾ method is applied in this study.

2.3. Physical Analysis of the Passive Film on 15 Mn Steel

AES (Auger Electron Spectroscopy) and XPS (X-ray Photoelectron Spectroscopy) were employed for surface analysis of the sample of 15 Mn steel. AES was performed with an accelerate voltage of 5 kV, a sample current of 0.1 μA, an experimental region of 10×10 μm, and a slope angle of 30 degree for the sample. In the case of the depth profile acquired by ASE, an argon spatter was used with a voltage of 2 kV, and a rate of 1 nm/min. XPS spectra were acquired by Al Kα of 600 μmφ for the sample of 15 Mn steel with take-off angles of 30, 60 and 90 degrees.

In order to investigate the surface (passive) film at the nano level, TEM was employed for the sample of 15 Mn steel. After the sample was mounted in resin, it was mirror-polished using emery paper and diamond paste. The passive film of 15 Mn steel was cut by an FIB (focused ion beam) from above the sample using SEM. EELS – TEM analysis was carried out to examine the chemical state of Cr, Si, Mn, Ni and Fe in the upper and lower passive films of 15 Mn steel.

2.4. EIS Measurements

In order to investigate the crevice corrosion behavior of 15 Mn steel, EIS measurements were conducted in low pH and high chloride solutions. After polishing the surface with Emery paper, electrodes were fabricated out of 15 Mn steel. Then, EIS measurements were performed in 15 mass% NaCl solution of various pHs. EIS test conditions employed a 2-electrode system with a measurement frequency range of 20 kHz to 1.0 mHz and an applied voltage of 10 mV.

3. Results and Discussion

3.1. Crevice Corrosion Resistance of 15 Mn Steel

The crevice corrosion resistance of 15 Mn steel was estimated in 15 mass% NaCl solution at 60°C for 4 weeks. Figure 1 shows the corrosion test results for both 15 Mn steel and SUS 430 steel. The surfaces of the front (F) and the reverse (R) sides of the samples were observed following rust removal. Crevice corrosion is clearly observable on the front and back of 15 Mn and 430 steels. Although there appears to be less corrosion of 15 Mn steel, it is difficult to quantitatively estimate the corrosion only by surface observation.

Fig. 1.

Crevice corrosion test results of the front (F) and reverse (R) side for 15 Mn and SUS 430 steel.

Figure 2 presents depth profiles of 15 Mn and SUS 430 steels using the laser microscope after crevice corrosion test for 3 and 4 weeks. Depth profiles for the samples were taken along the lines from 1A to 1B in the Figure. For 430 steel, the corrosion is spread over a wide area. Moreover, the maximum depth is very large (350 and 550 μm at 3 and 4 weeks, respectively). In the case of 15 Mn steel, although corrosion is similarly widely spread, the maximum depth is very low (120 and 174 μm at 3 and 4 weeks respectively). Thus, the crevice corrosion of 15 Mn steel is quantitatively identified to be less than SUS 430 steel by laser microscopy.

Fig. 2.

Depth profiles of 15 Mn steel and SUS 430 steel after crevice corrosion test for 3 and 4 weeks.

Figure 3 shows Gumble plots of corrosion depths for 15 Mn and SUS 430 steels after crevice corrosion test for 3 and 4 weeks. The vertical (y) axis is the scaleed Gumbel parameter, and the horizontal (x) axis is the corrosion depth. Corrosion depths for 430 steel at 3 and 4 weeks are very high. In addition, the slope (Δd/ΔY) of corrosion depth curve for 430 steel is large. On the other hand, the corrosion depths for15 Mn steel at 3 and 4 weeks are very small, and the slope of the corrosion depth curve is small.

Fig. 3.

Gumble plots of depths of 15 Mn steel and SUS 430 steel after crevice corrosion test for 3 and 4 weeks.

Figure 4 shows Gumble paramerters (α and λ) for the corrosion depths of 15 Mn and SUS 430 steels after crevice corrosion test for 3 and 4 weeks. λ is the mode of the corrosion depth, and α corresponds to the slope of the Gumbel distribution in Fig. 3. The λ of 430 steel increases over time, reaching more than 140 μm by 4 weeks. Besides, the α of 430 steel increases over time reaching high value by 4 weeks, which demonstrates that the crevice corrosion is very high for 430 steel. On the other hand, although the λ of 15 Mn steel increases over time, the value is low less than 100 μm at 4 weeks. Furthermore, the α of 15 Mn steel has a low value at 4 weeks. Thus, it is quantitatively found from the Gumbel parameters that the crevice corrosion of 15 Mn steel is very low.

Fig. 4.

Gumble paramerters of α and λ for the depths of 15 Mn steel and SUS 430 steel after crevice corrosion test for 3 and 4 weeks.

In order to examine the crevice corrosion of 15 Mn steel more detail, EIS measurements were conducted for steels in high chloride and low pH solutions. Figure 5 shows the results of EIS spectra of 15 Mn steel in 15% NaCl solution of various pHs. In high pH solution of pH 3 and pH 4, the two resistance components and one capacitance can be recognised in the spectrum. The impedance in the higher frequency region (Z_h) is likely related to the resistance of the solution. Because the impedance in the lower frequency region (R_l) is very high, it is considered to be the passive film resistance. Thus, the capacitance as shown in the spectrum is considered to be the film capacitance. On the other hand, in the lower pH solution of pH 1 and pH 2, as the impedance in the lower frequency region (R_l) is very low, it is considered to be the charge transfer resistance (R_ct). Therefore, the capacitance in the spectrum is considered to be the double layer capacitance. Thus, the passive film of 15 Mn steel is destroyed in the solution of low pH under 2.

Fig. 5.

EIS spectra of 15 Mn steel in 15% NaCl solution of various pHs.

Figure 6 displays impedances (Z_1mHz) at 1 mHz of 15 Mn and SUS 430 steels in 15% NaCl solution of various pHs. The Z_1mHz of 15 Mn steel decreases remarkably from pH 3 to pH 2, showing a change from the passive to active state. For SUS 430 steel, the situation is the same as for 15 Mn steel, namely, Z_1mHz of 430 steel decreases remarkably from pH 3 to pH 2. Thus, initiation of crevice corrosion is thought to be nearly the same for 15 Mn and 430 steels. In other words, the crevice corrosion behavior is thought to differ in the propagation stages.

Fig. 6.

Impedances at 1 mHz of 15 Mn steel and SUS 430 steel in 15% NaCl solution of various pHs.

3.2. Surface Analysis of the Passive Film on 15 Mn Steel

In order to identify the corrosion-preventing mechanism of 15 Mn steel, the passive film was estimated by surface analysis. Figure 7 shows a depth profile of the elements in the passive film of 15 Mn steel obtained by AES after 4 weeks of corrosion test. The concentrations (atomic %) of Fe, Mn, Si, Cr Ni and Oxygen are shown. From the profile of oxygen, the oxide film thickness is estimated to be almost 2.5 min of sputtering time. In this film, the concentrations of Cr and Mn are low at the surface, and they increase with the sputter time. Thus, Cr and Mn are enriched in the lower film from 1.2 to 2.5 min. of sputtering time. On the other hand, Si is enriched in the upper film, and there is no special enrichment of Ni.

Fig. 7.

Concentration profiles of the elements in the passive film on 15 Mn steel by AES.

Figure 8 shows the results of XPS analysis of the elements in the passive film on 15 Mn steel. The XPS spectra were acquired for the sample with take-off angles of 30, 60 and 90 degrees. For Fe 2p, the peak at 712 and 707 eV are visible, which indicates that Fe is detected both as the oxide and the metal state. For Cr 2p, there is a strong peak at 578 eV and a small peak at 575 eV, which indicates that Cr exists mainly as an oxide state. This is very similar to Si 2p, which has a strong peak at 102 eV and a small peak at 99 eV. For Mn 2p, there is a peak at 642 eV measured at 30 degree, and a peak at 640 eV measured at 90 degree, which indicates that Mn exists as an oxide at the surface and a metal state at the interface. This is almost the same as the case of Fe. On the other hand, in the case of Ni 2p, only the peak at 854 eV is sharp, which indicates that Ni exists primarily in the metal state. From these results of XPS, the passive film contains Fe, Mn, Cr and Si, and that Ni is enriched at the interface of the film and the metal. However, it was not possible to verify the chemical state of each element in the film clearly using the peak shift of the XPS spectrum.

Fig. 8.

XPS analysis for the elements in the passive film on 15 Mn steel.

In order to investigate the nano structure of the passive film on 15 Mn steel, an FIB-TEM analysis was conducted in Fig. 9. Three distinct regions are recognized: (1) steel, (2) lower film, and (3) upper film. The total thickness of the passive film is almost 20 nm. In later experiments, the structures of and differences between the lower and upper films were investigated in detail.

Fig. 9.

FIB-TEM analysis for the passive film on 15 Mn steel.

The chemical state of each element in the passive film was identified using EELS, as shown in Figs. 10, 11, 12. The positions of the EELS observations are the same as those in Fig. 9.

Fig. 10.

TEM EELS spectra of Cr and Si for the passive film on 15 Mn steel (1: steel, 2: lower film, 3: upper film).

Fig. 11.

TEM EELS spectra of Mn and Ni for the passive film on 15 Mn steel (1: steel, 2: lower film, 3: upper film).

Fig. 12.

TEM EELS spectra of Fe for the passive film on 15 Mn steel (1: steel, 2: lower film, 3: upper film).

The Cr-L EELS spectra of the passive film of 15 Mn steel are shown at left in Fig. 10. Spectrum 2 (lower film) shows peaks for Cr-L₃ and Cr-L₂ at 578 and 588 eV, respectively. Spectrum 3 (upper film) shows the peaks for Cr-L₃ and Cr-L₂ at 579 and 588 eV, respectively. The peak ratios (Cr-L₂/Cr-L₃) for Spectrum 2 and 3 are 0.7 and 0.9, respectively. From a previous paper,²¹⁾ Spectrum 2 (lower film) is similar to Cr (II) O, and the spectrum 3 (upper film) is similar to Cr (III) O. Thus, Cr is thought to exist mainly as Cr (II) oxide state in the lower film, and Cr (III) oxide state in the upper film of 15 Mn steel.

The Si-L EELS spectra in the passive film of 15 Mn steel are shown at right in Fig. 10. The spectrum 3 (upper film) has peaks of Si-L at 110 and 128 eV. From a previous paper,²²⁾ Spectrum 3 is similar to that of SiO₂. On the other hand, Spectrum 2 adopts a shape intermediate between Spectra 3 and 1. Since Spectrum 1 is the zero-valent metal state, Spectrum 2 is thought to indicate the Si (II) O state. Thus, Si is thought to exist primarily as Si (II) oxide state in the lower film, and Si (IV) oxide state in the upper film of 15 Mn steel.

The Mn-L EELS spectra of the passive film of 15 Mn steel are shown at left in Fig. 11. Spectra 2 and 3 show the peaks for Mn-L₃ and Mn-L₂ at 641 and 652 eV, respectively. The peak intensities of Mn-L₂ and Mn-L₃ for Spectra 2 and 3 are taken from Fig. 11. Then, the calculated intensity ratios of (Mn-L₂/Mn-L₃) for Spectra 2 and 3 are found to be 0.6 and 0.4, respectively. From a previous paper,²³⁾ the shapes of spectrum 2 and 3 are similar to that of Mn (III) O. Moreover, the peak ratios (Mn-L₂/Mn-L₃) for the chemicals of Mn (II) O and Mn (III) O are 0.23 and 0.40, respectively. Thus, Mn is thought to exist mainly as Mn (III) oxide state in the passive film of 15 Mn steel.

The Ni-L EELS spectra of the passive film of 15 Mn steel are shown at right in Fig. 11. Spectrum 2 (lower film) has small peaks for Ni-L₃ and Ni-L₂ at 854 and 872 eV, respectively. However, it is difficult to identify the peaks clearly in Spectrum 3 (upper film). From XPS measurements, Ni was shown to be enriched at the interface of the film and the metal. These EELS results confirm those of XPS. That is, Ni is enriched only at the interface of the film and metal, and there is less Ni in the upper film.

The Fe-L EELS spectra in the passive film of 15 Mn steel are shown in Fig. 12. The spectrum 2 (lower film) has peaks for Fe-L₃ and Fe-L₂ at 708 and 722 eV, respectively. Spectrum 3 (upper film) also has peaks for Fe-L₃ and Fe-L₂ at 709 and 724 eV, respectively. That is, the peaks for Fe-L₃ and Fe-L₂ in Spectrum 3 are shifted 1 and 2 eV higher than those in spectrum 2. From a previous paper,²⁴⁾ the Spectrum 2 (lower film) is similar to Fe (II) O, and Spectrum 3 (upper film) is similar to Fe (III) O. Thus, Fe is thought to exist mainly as Fe (II) oxide state in the lower film, and Fe (III) oxide state in the upper film of 15 Mn steel.

Although more work is needed to identify precisely, the chemical state of each element in the passive film of 15 Mn steel has been examined in this study.

3.3. Nano Structure of Passive Film of 15 Mn Steel

In the crevice corrosion tests, 15 Mn steel showed a higher corrosion resistance than SM. In particular, Gumble parameters (α and λ) for corrosion depths of 15 Mn steel were much smaller than those for SUS 430 steel after crevice corrosion test.

In EIS measurements, the impedances (Z_1mHz) at 1 mHz of 15 Mn steel decreased remarkably from pH 3 to pH 2, showing destruction of the passive film. As the tendency was the same in the case of SUS 430 steel, initiation of crevice corrosion was thought to be the same for the two steels. In other words, the large difference in Gumble parameters (α and λ) for the corrosion depths concerns penetration of crevice corrosion.

TEM-EELS measurements were conducted to identify the nano structure of the passive film on 15 Mn steel. The passive film consists of 2 layers of upper and lower film. The chemical states are found to be Cr (II), Si (II) and Mn (III) and in the lower film and Cr (III), Si (IV) and Mn (III) in the upper film. In addition, Ni is enriched at the interface between the metal and the film.

Thus, in the lower layer, the film is thought to contain mainly Cr (II) and Si (II) in Fe oxides. On the other hand, in the upper layer, the film is thought to contain mainly Cr (III) and Si (IV) in Fe oxides. As the crevice corrosion resistance of 15 Mn steel is thought to depend on the passive film itself at the initial period, Cr and Si are identified as the effective elements. In more detail, 15 Mn steel contains 10% Cr and 4% Si as alloying elements. In EIS results, the corrosion resistance of 15 Mn steel was almost the same as that of SUS 430 steel containing 16% Cr. Thus, Si is thought to increase the crevice corrosion resistance of 15 Mn steel.

Considering the propagation of crevice corrosion, Ni is thought to prevent penetration of the corrosion into the steel for an extended period of time. Ni was not enriched at the interface in AES measurement, because the sensitivity of AES was not high. However, it was confirmed from TEM and XPS that Ni was enriched at the interface. Specially, the enrichment of Ni was shown in XPS. Moreover, Ni is a noble metal as opposed to Fe. Thus, the Gumble parameters (α and λ) for corrosion depths of 15 Mn steel containing 8% Ni were much smaller than those for SUS 430 steel without Ni after crevice corrosion test. With respect to Mn, it is an austenite former like Ni, and Mn is enriched at the interface in AES experiment. As the diffusion of Ni or Mn in metal solid is difficult at the room temperature, they are supposed to be enriched at the lower film and the interface. Thus, Mn is thought to act like Ni and prevent the penetration of crevice corrosion to some extent effectively.

Thus, 15 Mn steel can maintain a passive film composed of effective elements in a crevice corrosion environment. As the concentration of Cl ions is high under the environment where vibration dampers are used at the coastal area, the crevice corrosion of them should be estimated for the crevice structure. Therefore, 15 Mn steel are very useful, which can have high corrosion resistance against crevice corrosion in a high chloride environment.

4. Conclusions

An Fe–Mn–Si–Cr–Ni steel (15 Mn; a shape memory alloy) showed higher corrosion resistance than carbon steel (SM) in crevice corrosion tests in 15 mass% NaCl solution at 60°C. In particular, Gumble parameters (α and λ) for the maximum corrosion depths of 15 Mn steel were much smaller than those for SUS 430 steel after crevice corrosion test.

In AES and XPS analysis, the passive film on 15 Mn steel was shown to contain Fe, Mn, Cr, Si and Ni. Based on TEM-EELS, the passive film consists of 2 layers with C (II), Si (II) and Mn (III) in the lower film and Cr (III), Si (IV) and Mn (III) in the upper film. Ni was enriched at the interface between the metal and the film. Cr and Si were thought to be effective early, and Ni was effective to prevent penetration of crevice corrosion. It was found that 15 Mn steel could maintain a passive film composed of effective elements in a crevice corrosion environment.

Acknowledgement

The author thanks Dr T. Sawaguchi who prepared the sample, and Dr K. Tsuzaki and Mr T. Maruyama for advice concerning 15 Mn steel used in this study.

References

• 1)   L. C.  Chang and  T. A.  Read: Trans. Am. Inst. Min. Metall. Pet. Eng., 189 (1951), 47.
• 2)   S.  Miyazaki and  K.  Otsuka: ISIJ Int., 29 (1989), 353.
• 3)   K.  Otsuka,  H.  Sakamoto and  K.  Shimizu: Acta Metall., 27 (1979), 585.
• 4)   A.  Sato,  E.  Chishima,  K.  Soma and  T.  Mori: Acta Metall., 30 (1982), 1177.
• 5)   A.  Sato and  T.  Mori: Sci. Eng. A, 146 (1991), 197.
• 6)   H.  Otsuka,  H.  Yamada,  T.  Maruyama,  H.  Tanahashi,  S.  Matsuda and  M.  Murakami: ISIJ Int., 30 (1990), 674.
• 7)   H.  Fukai,  S.  Suzuki,  N.  Masashi,  S.  Hanada,  T.  Maruyama,  H.  Kubo and  Y.  Waseda: Mater. Trans., 46 (2005), 1745.
• 8)   T.  Sawaguchi: Bull. Iron Steel Inst. Jpn., 20 (2015), 96.
• 9)   T.  Nishimura: Mater. Trans., 55 (2014), 871.
• 10)   T.  Shibata: ISIJ Int., 31 (1991), 115.
• 11)   A.  Tahara and  T.  Shinohara: Corros. Sci., 47 (2005), 2589.
• 12)   N.  Taniguchi,  H.  Suzuki,  M.  Kawasaki,  M.  Naito,  M.  Kobayashi,  R.  Takahashi and  H.  Sano: Corros. Eng. Sci. Technol., 46 (2011), 117.
• 13)   P.  Schmutz,  V.  Guillaumin,  R. S.  Lillard,  J. A.  Lillard and  G. S.  Frankel: J. Electrochem. Soc., 150 (2003), No. 4, B99.
• 14)   C. M.  Oertel,  S. P.  Baker,  A.  Niklasson,  L.-G.  Johansson and  J.-E.  Svensson: J. Electrochem. Soc., 156 (2009), No. 12, C414.
• 15)   K.  Sapre,  S.  Seal,  P.  Jepson,  Z.  Rahman and  T.  Smith: Corros. Sci., 45 (2003), No. 1, 59.
• 16)   M.  Reid,  J.  Punch,  L. F.  Garfias-Mesias,  S.  Belochapkine and  D. A.  Tanner: J. Electrochem. Soc., 155 (2008), No. 4, C147.
• 17)   M.  Reid,  J.  Punch,  G.  Grace,  L. F.  Garfias and  S.  Belochapkine: J. Electrochem. Soc., 153 (2006), No. 12, B513.
• 18)   T.  Nishimura: Corros. Sci., 52 (2010), 3609.
• 19)   T.  Nishimura,  H.  Katayama,  K.  Noda and  T.  Kodama: Corrosion, 56 (2000), 935.
• 20)   J. S.  White: Indian Math., 14 (1964), 21.
• 21)   T. L.  Dailton,  B.  Little and  K.  Lowe: Ultramicroscopy, 106 (2006), 561.
• 22)   T.  Nishimura: Mater. Trans., 48 (2007), 1438.
• 23)   L.  Laffont and  P.  Gibot: Mater. Charact., 61 (2010), 1268.
• 24)   H.  Tan,  J.  Verbeeck,  A.  Abakumov and  G.  Van Tendeloo: Ultramicroscopy, 116 (2012), 24.