Peng-Lin Gao(高朋林) Jian Gong(龔建) Qiang Tian(田強) Gung-Ai Sun(孫光愛) Hai-Yang Yan(閆海洋)Liang Chen(陳良) Liang-Fei Bai(白亮飛) Zhi-Meng Guo(郭志猛) and Xin Ju(巨新)

1Department of Physics,University of Science and Technology Beijing,Beijing 100083,China

2Key Laboratory of Neutron Physics and Institute of Nuclear Physics and Chemistry,China Academy of Engineering Physics,Mianyang 621900,China

3Institute of Power Metallurgy,University of Science and Technology Beijing,Beijing 100083,China

Keywords: oxide dispersion strengthened(ODS)steel,small angle neutron scattering(SANS),thermal aging,nanoparticle

1. Introduction

Oxide dispersion strengthened (ODS) steels are considered to be one of the most promising candidate structural materials for generation IV fission and fusion reactors.[1–7]Their attractiveness as nuclear structural materials lies in their excellent irradiation tolerance and creep strength at high temperature.[3–6]In comparison with conventional steels,ODS steels contain a high density of oxide nanoparticles.[7,8]These oxide nanoparticles block dislocation motion and grain boundary migration,and thus effectively increase the ultimate tensile strength and almost double the yield strength compared with conventional steels above 773 K.Therefore,the thermal stability of oxide nanoparticles in ODS steels is an important factor that directs their application.

Oksiutaet al.[9]and Zhonget al.[10]observed that the Y–Ti–O-enriched nanoparticles in 14Cr-ODS steel exhibited high stability at several annealing temperatures below 1573 K for 1 h and decreased micro-hardness above 1573 K for 1 h due to significant coarsening of the nanoparticles. Ribiset al.[11]suggested that the high stability of nanoparticles might be the result of a low interfacial energy. It was reported[12–14]that the Cr-richα′phase forms in 20Cr-ODS steel and the micro-hardness is significantly increased during thermal aging at 758 K for 1000 h. Although many studies have reported on the effects of thermal aging on dispersed nanoparticles in high-Cr ODS steels,the stability of nanoparticles in 9Cr-ODS steel during long-term thermal aging has not been extensively studied.

The small-angle neutron scattering (SANS) technique is a powerful tool for probing nano-sized precipitation in ferromagnetic steels.[15–18]Because neutrons carry no charge they can penetrate deeply into bulk materials and detect buried structures. Further, neutrons have a small magnetic moment and are sensitive to magnetic structures. ODS steels can be considered as non-magnetic nanoparticles dispersed in a ferromagnetic matrix. SANS is the most convenient tool for studying this dispersion. When a saturating magnetic field is applied to the sample, statistically quantitative information about the size, shape, volume fraction and number density of the precipitate can be obtained.[19]Furthermore,additional information on the composition of the precipitates can be obtained from the ratio of magnetic scattering to nuclear scattering intensities.[14]

As reported in several papers,[20–22]the typical operating temperatures for ODS steels range from 823 K to 973 K.Throughout this work, the thermal aging temperature was set at 873 K.The stability of oxide nanoparticles in 9Cr-ODS steel during long-term thermal aging treatments up to 5000 h was studied by SANS, transmission electron microscopy (TEM)and Vickers micro-hardness measurements.

2. Experimental procedure

The 9Cr-ODS steel was prepared by mechanical alloying of a mixture of the corresponding metallic powders with Y2O3.Then the resulting mixture was consolidated by hot isostatic pressing(HIP)at 130 MPa for 4 h at 1443 K.The steel was normalized at 1323 K for 1 h followed by air-cooling,and was subsequently tempered at 1053 K for 1 h followed by aircooling. The chemical composition of the steel used in this study is given in Table 1. Isothermal aging at 873 K was performed on the 9Cr-ODS steel for durations of 100 h, 500 h,1000 h,3000 h,and 5000 h under ambient pressure.

Table 1. Chemical composition of 9Cr-ODS steel(wt.%).

SANS measurements of the samples were carried out at the China Mianyang Research Reactor on the Suanni smallangle neutron spectrometer. The measurements were performed using neutron wavelength(λ)of 0.53 nm and the distances between the sample and the detector were 1.9 m and 5.9 m,respectively.A scattering vector(q)range of 0.08 nm-1to 2 nm-1was covered.During measurements,saturated magnetic fields (H=1.5 T) perpendicular to the direction of the incident neutron beam were applied to the samples. The azimuthal dependence of scattering intensities is expressed as[23]

whereInucandImagare the nuclear and magnetic scattering intensity, respectively, andαis the angle between the magnetic field direction and scattering vectorq. TheI–qcurves were radially averaged by taking 20°sectors in the horizontal(α=0)and vertical(α=90)directions of the detector plane.According to Eq.(1),the magnetic scattering intensities were calculated byImag=I(α=π/2)-I(α=0). The measured data were corrected for sample transmission,background and detector efficiency. The absolute intensity was calibrated using a 1 mm thick water sample. Data reduction was processed using the program BerSANS.[24]

The TEM experiments were performed with a FEI Tecnai G20 FEG microscope at an accelerating voltage of 200 kV.Thin foils for TEM observation were mechanically polished to a thickness of about 100 μm and then electropolished by the twin-jet technique using an electrolyte consisting of 95 vol.%ethanol and 5 vol.%perchloric acid at 233 K.Electron backscattered diffraction(EBSD)analysis was performed in a field emission scanning electron microscope with a step size of 0.12 μm. The micro-hardness of these specimens was measured with a Vickers diamond pyramid under a load of 9.8 N for 10 s. Each micro-hardness value was determined through averaging 10 independent measurements of the same sample.

3. Results and discussion

3.1. Model

The ODS steel under study can be considered as a dispersion of non-magnetic spherical oxide particles in a ferromagnetic matrix. The magnetic scattering intensity from a system of polydisperse particles can be described as[25,26]

whereRmis the median radius,σis the width of the size distribution andNis the number density. The mean radius of nanoparticles is evaluated asRmexp(σ2/2). The distribution of the volume fraction of the nanoparticles isDv(R)=N(R)Vp(R).

The SANS magnetic scattering curves of the aged 9Cr-ODS samples are shown in Fig. 1(a). No obvious changes can be observed forqranging from 0.3 nm-1to 2 nm-1,which corresponds to the scattering from the oxide nanoparticles. Asqapproaches 0, the form factorF(qR) is close to 1, and thusI(q) should become constant, which can be derived from Eqs. (2)–(5). However, as shown in Fig. 1(a),the measured magnetic scattering intensities increased at lowqvalues, due to the scattering from large inhomogeneities at length scales larger than 30 nm(as estimated byπ/q)such as carbides and magnetic domain boundaries. The low-qregion(q ＜0.2 nm-1) could be fitted by a power law[27]which decays very fast toward highq. Therefore, the SANS magnetic scattering data in the measured fullqrange were modelled by

Fig. 1. (a) SANS magnetic scattering intensities for 9Cr-ODS steel for unaged sample and five different aging durations. (b) Measured (circles) and fitted(line)SANS spectra of the unaged sample.

whereCis a constant that relates to the background. An example of a fitted profile is shown in Fig. 1(b). All the fitted curves matched well with the measured data.TEM images were analyzed using the program Nano Measurer to obtain the nanoparticle size distribution; more than 1000 particles in each specimen were statistically analyzed.Figure 4 shows the size distributions obtained by TEM and SANS techniques for the unaged and aged samples. The results are in good agreement with each other within the uncertainty of the measurements for each sample. Since the scattering intensity is proportional to the sixth power of particle size,the SANS data were weighted by the larger particles. Accordingly,the average particle size obtained from SANS is slightly larger than that derived from the TEM results.

3.2. Effects of thermal aging on oxide nanoparticles

The size distribution of oxide nanoparticles in 9Cr-ODS steel under various aging conditions is described pictorially in Fig.2. The calculated magnetic contrast of 9Cr-ODS steel was 4.5×1010cm-2. The characteristics of the mean size of nanoparticles and the volume fractions are listed in Table 2.The mean radius of the precipitates of the unaged sample was 2.54±0.05 nm,with a volume fraction of 0.57±0.03 vol.%.For the sample aged for 5000 h, the mean radius and volume fraction of the nanoparticles were 2.52±0.05 nm and 0.53±0.03 vol.%, respectively. Within the uncertainties of the measurements, the fitted size distributions of all the aged samples remained unchanged.

Table 2. Characteristics of nanoparticles deduced from SANS.

TEM analyses were performed to cross-check the SANS results for the samples.TEM micrographs of the aged samples are shown in Fig.3. They indicate that the oxide nanoparticles were homogenously distributed in the matrix. In this study,

Fig.2. Nanoparticle size distributions in thermally aged 9Cr-ODS steel.

Fig. 3. TEM micrographs of 9Cr-ODS steel (a) unaged; and (b) 100 h, (c)500 h,(d)1000 h,(e)3000 h,(f)5000 h thermally aged samples.

TheAratio is defined as the ratio of normalized scattering intensities perpendicular and parallel to the magnetic field direction[28]

Fig.4. Size distributions obtained by TEM and SANS for(a)unaged and(b)100 h,(c)500 h,(d)1000 h,(e)3000 h,and(f)5000 h aged samples.

Based on the SANS study by Mathonet al.,[19]the nanoparticle composition for 9Cr-ODS was assumed to be Y2Ti2O7and the theoretical value of nuclear contrast is 3.66×1010cm-2. TheAratio for Y2Ti2O7calculated from Eq. (7)was estimated to be 2.5, in good agreement with the ratios obtained from our measurements. As shown in Table 2, theAratio varies around 2.5 and keeps stable within the errors.Therefore, the chemical composition of the particles was not affected by the thermal treatments.

The micro-hardness of the aged samples as a function of the thermal aging time is shown in Fig.5. The micro-hardness of the unaged sample was 365 Hv. The micro-hardness of the aged samples was found to be similar to that of the unaged samples. This relates to the high thermal stability of the extremely fine oxide nanoparticles, as evidenced by the SANS magnetic scattering and TEM results. Representative inverse pole figure(IPF)maps and the grain size distribution obtained by analyzing EBSD results of specimens aged for different times are shown in Fig.6. The IPF maps of 9Cr-ODS steel before aging(Fig.6(a))and after aging up to 5000 h(Fig.6(b))revealed almost no change in grain structure. The average grain size before and after the aging process at 973 K for 5000 h were about 0.9 μm and 1.0 μm,respectively.The grain sizes exhibited only a slight change. This could be related to an even distribution of oxide nanoparticles,which acted to inhibit the migration of grain boundaries and dislocations during thermal aging.

In this study,the Y–Ti–O nanoparticles dispersed in 9Cr-ODS steels exhibited good stability at 873 K.In contrast to our recent work on as-milled mechanical alloyed 9Cr-ODS powders, the precipitation of nanoparticles with a radius of about 1 nm was observed by in situ SANS measurements at 873 K.In that study,prolonged annealing time or increasing temperature led to coarsening of the nanoparticles.[29]That is, the oxide nanoparticles could precipitate in the milled ODS powder at～873 K.However, in 9Cr-ODS steel prepared by HIP at 1443 K,the nanoparticles remained stable at lower temperatures. Oonoet al.[30]also found that the nanoparticles in 9Cr-ODS steel began to coarsen at about 1473 K during heating for 200 h, and had excellent thermal stability below 1473 K,in agreement with this work. It can be remarked here that the operating temperatures ranged from 823 K to 973 K,hence the prepared 9Cr-ODS is a potential candidate structural material for generation IV fission and fusion reactors.

Fig.5. Evolution of the Vickers micro-hardness measured on samples after different thermal aging treatments.

Fig.6.Inverse pole figure maps and grain size distribution obtained by EBSD analyses for(a),(b)unaged and(c),(d)5000 h aged samples.

The high stability of 9Cr-ODS steel is attributed to two factors. First, there is a coherent relationship between the Y–Ti–O oxide nanoparticles and the matrix, resulting in a low interfacial energy and high stability of the oxide nanoparticles.[11,31,32]Secondly, the structural stability is dependent on the Cr content in the steel. Ferritic and duplex stainless steels with a Cr content above～12 wt.%are known to harden and embrittle after thermal aging at temperatures between 673 K and 823 K due to the formation of a Cr-richα′phase.[14,22]In this study, the 9Cr-ODS steel with a low Cr content of 8.88 wt.%showed good structural and performance stability.

In summary,the Y–Ti–O oxide nanoparticles in 9Cr-ODS steel exhibited good thermal stability after aging for 5000 h at 873 K.These nanoparticles played a key role in preventing grain growth and dislocation motion, and therefore conferred good thermal stability on the prepared 9Cr-ODS steel. The structural evolution of the Y–Ti–O nanoparticles in ODS steel at higher temperatures and longer aging times is a desirable goal for future study.

4. Conclusion

The stability of the oxide nanoparticles in 9Cr-ODS steel during thermal aging at 873 K for up to 5000 h was investigated by SANS and TEM analysis. The SANS technique showed that the size distributions of oxide nanoparticles did not change obviously after long-term thermal aging. The mean radius and volume fraction of the oxide nanoparticles in the thermally aged sample were estimated to be 2.50 nm and 0.55 vol.%,respectively. These values are in good agreement with the results obtained from TEM observations. Furthermore,no significant changes to micro-hardness and grain size were found after aging, which can be attributed to the good thermal stability of the oxide nanoparticles.

Acknowledgment

Project supported by the National Key Research and Development Program of China(Grant No.2017YFB0702400).