Phase Transformation and Mechanical Properties of G Phase (Mn6Ni16Si7) in Mn–Ni–Si Model Alloys after 1,000°C Annealing

Xinrun Chen; Tatsuya Suzuki; Huilong Yang; Ba Vu Chinh Nguyen; Zhehuan Zhang; Kenta Murakami

doi:10.2320/matertrans.MT-M2024052

Abstract

Mn₆Ni₁₆Si₇ intermetallic compounds, referred to as the G phase, and their precursors are recognized as potential secondary phases precipitated in neutron-irradiated steels and are the main cause of embrittlement in low-alloyed reactor pressure vessel steels under long-term operation. To obtain the mechanical properties and thermal stability of the G phase, two Mn–Ni–Si model alloys (composed of 21 mol%Mn-58 mol%Ni-21 mol%Si and 30 mol%Mn-58 mol%Ni-12 mol%Si) were annealed at 1,000°C. The existence of the G phase in both annealed ternary model alloys was confirmed by XRD and SEM/EDS. Meanwhile, the process of different Mn–Ni–Si ternary regions to the G phase transformation under 1,000°C annealing has also been discussed. Young’s modulus and nano-hardness of the G phase were measured using the nanoindentation technique. The results showed similar values for the G phase in two alloys with Young’s modulus of approximately 220 GPa, which is similar to Young’s modulus of iron. This fact suggests the necessity of reconsideration of the hardening model contributed by G phase precipitates in reactor pressure vessels in the future.

1. Introduction

The embrittlement of reactor pressure vessel (RPV) steels is highly related to the long-term operation of nuclear power plants. Recently, in both high-Cu steels and low-Cu alloyed steels under long-term operation (approximately 40 years) at approximately 300°C, nano-scale Mn–Ni–Si (MNS) clusters were observed using microstructure techniques [1]. Furthermore, MNS clusters and their precursors are the main cause of embrittlement of RPV materials under exposure to a high flux of neutron irradiation from the adjacent core. Recently, Mn₆Ni₁₆Si₇ intermetallic compounds [2–4], namely, the G phase, and its precursors [5], are considered candidates for the secondary phases precipitated in neutron-irradiated ferritic steels.

In addition to its importance in the nuclear field, the G phase (Mn₆Ni₁₆Si₇) with Mg₆Cu₁₆Si₇-type structure [6–8] has been studied in diverse fields owing to its excellent magnetic inhibition properties and superconductivity. The refined crystal structure of Mn₆Ni₁₆Si₇ was reinvestigated by Yan et al. using a single crystal combined with X-ray and neutron powder diffraction [9]. Moreover, King et al. predicted the preferential condition of the B2-G phase structure transformation in a Fe matrix using a density functional theory simulation [10]. Furthermore, for the MNS ternary system, including the G phase, an isothermal section at 800 and 1,000°C of MNS system was summarized by K.P. Gupta [11]. In addition, Hu et al. performed thermodynamic modeling of the MNS system based on previous and experimental data at 800 and 1,000°C, and improved the isothermal section at 1,000°C of the MNS system [12].

Recently, to improve the degradation prediction model for evaluating the integrity of RPV after long-term operation, the mechanical properties, formation mechanisms, and information on the precursors of the G phase have become essential for improving the hardening model contributed by the G phase [1, 2]. However, to the best of our knowledge, some mechanical properties, such as Young’s modulus, are still unknown. Therefore, this study aims to: (1) successfully synthesize a sufficient size of G phase in two microscale polycrystalline MNS ternary model alloys through annealing, with the stoichiometric composition similar to the G phase as the main body and a higher Mn one for comparison; (2) identify the ternary phases before and after annealing and evaluate the process of G-phase transformation using XRD and SEM/EDS; and (3) accurately measure Young’s modulus and hardness of the G phase in microscale polycrystalline MNS ternary model alloys using nanoindentation and advise on improving the hardening model.

2. Experimental Procedure

2.1 Process and materials

Two MNS ternary model alloys weighing approximately 20 g were fabricated by melting a mixture of pure manganese (99.9 mass%), nickel (99.96 mass%), and silicon (99.99 mass%). The stoichiometric compositions were set as 21 mol%Mn-58 mol%Ni-21 mol%Si and 30 mol%Mn-58 mol%Ni-12 mol%Si, referred to as the 21Mn-58Ni-21Si and 30Mn-58Ni-12Si. The ratio of the 21 mol%Mn-58 mol%Ni-21 mol%Si alloy was approximately the same as the G phase such that it is referred to as Specimen-G. The ratio of the 30 mol%Mn-58 mol%Ni-12 mol%Si alloy was selected to compare with a previous study such that it is referred to as Specimen-M [12]. Raw materials were weighed using a high-precision balance with a sensitivity of 0.0001 g. Cylindrical samples were fabricated in a vacuum-induced melting furnace at 1,430°C and for 3 min to improve their homogeneities under a 99.996 vol% pure Ar atmosphere. According to previous studies [12], approximately 1 mass% of additional manganese was added to the weighed materials to compensate for the Mn loss during vacuum-induced melting owing to the volatilization of Mn at the over-melted point temperature. Subsequently, the cylindrical samples were cut into plates with dimensions of 5 × 8 × 0.7 mm³ using electrical discharge machining. To prepare a good surface for subsequent chemical and crystal analysis, as well as nanoindentation hardness measurement, the alloys after melting were well rough-polished by SiC polishing paper in the order of 46, 18, and 5 µm, and were fine-polished using 1 µm diamond solution and 0.04 µm SiC solution on a soft pad.

2.2 Annealing tests

The alloys after melting were checked for chemical analysis using SEM/EDS, and the difference in the chemical composition of each sample with the designated stoichiometric composition was less than 2 mol%.

Subsequently, plates of both alloys were covered with Mo film and annealed at 1,000 ± 2°C at approximately 10⁻³ Pa for 4 days in a sealed and vacuumed single closed quartz tube of annealing furnace equipped with a molecular pump and thermocouples followed by water quenching.

2.3 XRD and SEM/EDS analysis

Five parts of each plate before and after annealing were checked for chemical analysis using a scanning electron microscope equipped with energy dispersive X-ray spectroscopy system (SEM/EDS, JEOL JCM-6000Plus/EDS) in a second electron imaging pattern of SEM and mapping analysis pattern of EDS in a rectangular shape area of 800 × 600 µm². The chemical composition of each phase in the ternary alloys was checked by point analysis of approximately 30 points of each phase in an area of 200 × 150 µm². The mean values of the mapping and point analyses were then selected to evaluate the chemical compositions of the ternary alloys and related phases. In addition, the area change of the phases before and after annealing was calculated in five parts of each plate using binarized foreground and background methods using the MATLAB graphics toolbox functions.

The information on the crystal structure of each specimen was obtained using an X-ray diffractometer (XRD, Miniflex 600) with Cu Kα radiation 40 kV and 15 mA with a high-speed 1D detector D/teX Ultra2 at a scanning speed of 3 deg/min. Each plate was measured in three different directions to reduce the effect of the grain crystal orientation. The data obtained were matched to the ICDD-PDF2 database.

2.4 Nanoindentation tests

The hardness and Young’s modulus of the annealed phases were measured using a nanoindentation machine (Shimadzu DUH-211S) with a Berkovich indenter and a load-unloading method at a rate of 1.4 mN/s. A maximum indentation depth of 300 nm was set to minimize the impact of the work-hardened layer. Each indentation was selected at the center of each phase to further evaluate the effective coverage of nano-hardness and Young’s modulus; approximately 100 valid indentations per phase were obtained despite scattering.

3. Results and Discussions

3.1 Crystal structure identification and phase transition discussion

Two regions were observed in Specimen-G after melting, as shown in Fig. 1. One was a region with a composition close to the target composition and occupied approximately 80 vol% of the sample. Most of the grain size in this region was larger than 100 µm in diameter; however, some of the grains showed a solidification structure that intermingled with the other region in a dendritic manner, with a diameter of approximately 5 µm. The second region had a high Ni concentration of approximately 75 mol% and was likely to be the Ni₃X (X = Mn, Si) intermetallic compound. The grain size was approximately 5 µm and showed a dendritic structure together with the first region. SEM and EDS images of this sample after annealing at 1,000°C for 4 days showed a slight decrease in Mn concentration from the original Specimen-G. The Mn concentration decreases during vacuum annealing. The ratios of the two regions remained nearly the same after annealing. However, the intricate microstructure showed that the G phase-like grains were more rounded and encapsulated by a higher Ni phase. At first glance, Specimen-G did not appear to have changed much before and after annealing, and the ratio of the two regions remained nearly the same. However, as shown in Fig. 2, the XRD results indicate that an annealing-induced phase transformation occurred. Cubic Pm3m(221) structures were identified in the sample immediately after melting, corresponding to Ni₃Si and MnNi₃. However, despite the existence of a large area with a chemical composition similar to that of the G phase, XRD peaks of the G phase were not observed. After annealing, significant XRD peaks corresponding to the G-phase were observed. As summarized in Table 1, the chemical compositions of each area before and after annealing did not differ significantly. Therefore, it is suggested that only a region similar in composition but not in structure to the G phase was mostly present in the dissolution and production stages of sample G and it underwent a phase transformation to the G phase after annealing.

Fig. 1

SEM/EDS analysis of Specimen-G before and after annealing (the unit of composition is mol% in the Mn-Ni-Si region).

Fig. 2

XRD results of Specimen-G before and after annealing.

Table 1 SEM/EDS analysis of chemical composition before and after annealing.

Two or more regions were also observed in Specimen-M, as shown in Fig. 3. Approximately 60 vol% of the samples are regions with a composition of 33 mol%Mn-55 mol%Ni-12 mol%Si, referred to as 33Mn-55Ni-12Si, and their grain sizes exceed 50 µm. A region with a dendritic structure, which suggests the formation of Ni₃X (X = Mn, Si), was also found, with slight heterogeneity of Mn in this region. Notably, as shown in Fig. 4, XRD peaks of the G phase were also not observed before annealing. The XRD results suggest that the crystalline phase is mostly cubic Pm3m(221), which is the same trend as in Specimen-G. After annealing, the concentration of Mn in the overall sample decreased, whereas those of nickel and silicon increased. The area with a high Ni concentration increased from approximately 40 vol% to 60 vol%. The grains with a composition of 33Mn-55Ni-12Si disappeared, and a new phase with a concentration similar to that of the G phase was formed. The XRD pattern shows a mixture of the G-phase and Ni₃X. Therefore, Specimen-M also demonstrates G-phase transformation despite the significant differences in the original chemical composition.

Fig. 3

SEM/EDS analysis of Specimen-M before and after annealing (the unit of composition is mol% in the Mn-Ni-Si region).

Fig. 4

XRD results of Specimen-M before and after annealing.

It is clear from Table 1 and Fig. 5 that a direct comparison of the annealing in the two samples is not appropriate. Notably, Mn has a high equilibrium vapor pressure and is therefore considered unstable in the MNS ternary system at 1,000°C under vacuum conditions. In Specimen-G, the overall chemical composition of the specimen, ratio of the phase area, and chemical composition in the region of G phase transformation changed slightly. The difference in the overall chemical composition is owing to the evaporation of Mn during annealing, suggesting the high thermal stability of the Mn solute in the structural transformation of the G phase. Slight changes in the G phase transformation region are achieved with minute changes in solutes by diffusion during the G phase structural transformation, demonstrating the high stability of Mn in this case from the side as well. However, in Specimen-M, the overall chemical composition, ratio of the phase area, and chemical composition in the region of G phase transformation changed significantly. This suggests that a complicated phase transformation occurs during annealing, accompanied by the evaporation of unstable Mn. One assumption is that the G-phase transformation, in this case, could have proceeded from a high-Mn-concentration MNS ternary region through a process consisting of significant chemical composition transformation of Mn loss and structural transformation. However, it is interesting to note that in both cases, the G phase did not form immediately after melting, regardless of whether the composition in the majority of the samples is similar to the G phase or not, and underwent annealing to the G phase.

Fig. 5

Histogram of phase area changing in two specimens before and after annealing.

3.2 Nano-hardness and Young’s modulus of G phase

Table 2 summarizes the typical values of indentation hardness and Young’s modulus distribution of different phases in the two microscale polycrystalline MNS ternary model alloys. Despite slight differences in composition and grain size, the nanoindentations in the G-phase-like regions of the two specimens showed similar values, with Young’s modulus of approximately 220 GPa. The hardness and Young’s modulus distributions were nearly symmetrical, with the difference between the average and median values of the hardness distribution and Young’s modulus being less than 0.1 and 0.9 GPa, respectively. The distribution of hardness and Young’s modulus obtained by the above procedures represents the global features of the nano-hardness and Young’s modulus of the G phase in the polycrystalline MNS ternary model alloys. The difference in hardness and Young’s modulus between the two alloys was caused by the difference in the grain sizes of the G phase in the polycrystalline MNS ternary model alloys. Because the difference in Young’s modulus is less than 3% for each model alloy, the average value of approximately 220 GPa can be considered as the typical value of Young’s modulus for the G phase (Mn₆Ni₁₆Si₇) on a microscale.

Table 2 Typical values of hardness distribution and Young’s modulus distribution of different phases in two model alloys.

Concerning the hardening model for the assessment of the integrity of the RPV after a long-term operation, because the G phase has a high hardness and Young’s modulus similar to that of iron (approximately 210 GPa), it is essential to reconsider the hardening model contributed by G-phase precipitates in recently applied reactor pressure vessel steels. Russell and Brown described a hardening model in which the difference in Young’s modulus affects the curvature of dislocations based on the fact that the Cu precipitates in Fe–Cu alloys are smaller than the Fe matrix in both hardness and Young’s modulus [13]. This method has been applied to old reactor pressure vessels, where Cu-rich clusters/precipitates were the main contributors to embrittlement. Even in MNS clusters, Young’s modulus is smaller than that of the Fe matrix if the structure is body-centered cubic and the Fe concentration in the cluster is sufficiently high. However, in newer reactor pressure vessels of low-alloyed steels, hard precipitates, such as the G phase, may be more appropriate as the main contributor to irradiation hardening by inhibiting the movement of dislocations [14]. Therefore, the precise Young’s modulus of these hard precipitates is essential for the improvement of the hardening model.

4. Conclusion

In this study, the sufficient size of the G phase (Mn₆Ni₁₆Si₇) was successfully formed in two microscale polycrystalline Mn–Ni–Si ternary model alloys after 1,000°C annealing. The nano-hardness and Young’s modulus of the G phase were measured, and the phase transformation was experimentally discussed. The conclusions are as follows:

(1) In Specimen-G, therefore, it is suggested that only a region similar in composition but not in structure to the G phase was mostly present in the dissolution and production stage, and then it underwent a phase transformation mainly in the structural transformation to the G phase after annealing accompanied by slight changes in the chemical composition.
(2) Specimen-M demonstrated G-phase transformation despite the significant differences in the original chemical composition before and after annealing. This suggests that the G-phase transformation, in this case, could have proceeded from a high-Mn-concentration MNS ternary region through a process consisting of a significant chemical composition transformation of Mn loss and a structural transformation.
(3) Despite slight differences in composition and grain size, the nanoindentations in the G-phase (Mn₆Ni₁₆Si₇)-like regions of the two specimens showed similar values, with Young’s modulus of approximately 220 GPa, which is similar to that of iron. This suggests the necessity of reconsidering the hardening model contributed by G-phase precipitates in recent reactor pressure vessels.

Acknowledgments

A Part of this study was supported by JSPS KAKENHI Grant Number 21H01857 and MEXT Innovative Nuclear Research and Development Program Grant Number 23813718.

The SEM/EDS observations and XRD ICDD-PDF2 database matching were performed at the Nagaoka University of Technology Analysis and Instrumentation Center. We thank Nagaoka University of Technology Analysis and Instrumentation Center for use of facilities and equipment.

The XRD measurements were performed at the Nagaoka University of Technology Radioisotope Center. We thank Nagaoka University of Technology Radioisotope Center for use of facilities and equipment.

REFERENCES

Corresponding author

Register with J-STAGE for free!