Materials Transactions, Vol. 55, No. 3 (2014) pp. 423 to 427 Special Issue on In Situ TEM Observation of High Energy Beam Irradiation © 2014 The Japan Institute of Metals and Materials
New Approach to In Situ Observation Experiments under Irradiation in High Voltage Electron Microscopes Hiroaki Abe1,+1, Takahiro Ishizaki2,+2, Feng Li2, Sho Kano2, Yanfen Li1, Yuhki Satoh1, Takeshi Nagase3 and Hidehiro Yasuda3 1
Institute for Materials Research, Tohoku University, Sendai 980-8577, Japan Graduate School of Engineering, Tohoku University, Sendai 980-8579, Japan 3 Research Center for Ultra-High Voltage Electron Microscopy, Osaka University, Ibaraki 567-0047, Japan 2
A new method for conducting in situ observations of experiments undergoing irradiation in a high voltage electron microscope (HVEM) is proposed. Intensity profile of a focused electron beam in HVEM introduces an atomic displacement gradient in the vicinity of the beam, which generates distribution of point defect concentration and enhances defect diffusion in matrix. In our experiments, tantalum carbide or yttrium titanate nanometer-scale particles embedded in iron matrices were irradiated at 673 K with a focused electron beam at energy ranges from 0.75 to 2.5 MeV. The results show that the instabilities of particles undergoing irradiation could be observed as diminishing either in size or contrast. The rate of shrinkage per fluence unit was successfully measured to derive the vacancy diffusion effect, with particles located in the vicinity of the electron beam showing higher rates of shrinkage. This indicates that the diffusion of vacancies enhanced both by irradiation and the concentration gradient is attributable to dissolution of the particle constituents into the matrix. [doi:10.2320/matertrans.MD201315] (Received October 1, 2013; Accepted November 20, 2013; Published January 18, 2014) Keywords: high voltage electron microscope, electron irradiation, in situ observation, tantalum carbide, yttria titanate, iron, F82H steel, oxide dispersion strengthened steel
1.
Introduction
Radiation-induced defects and their kinetics were first developed in high voltage electron microscopes (HVEMs) in the 1960s,1) and applied to metals,1,2) alloys,35) and ceramics.6) The fundamental kinetic equations were developed as described in Refs. 1,711). Further development was made towards ion irradiation and neutron irradiation.12) Important insights derived from the systematic investigations were the activation energy of interstitial atom migration,8) radiationenhanced diffusion,13) radiation-induced diffusion,9) displacement threshold energy,14) and the effect of the displacement cascade,15) and others. Electron irradiation experiments in HVEMs have an advantage over electron microscopes in that adjusting the degree of beam focus can produce beam intensity variations. Beam intensity distributions are close to Gaussian or CauchyLorentz distributions under focused conditions. In contrast, those distributions assume a bell jar distribution under defocused conditions. Under normal circumstances, researchers employ properly defocused beams in order to achieve the same intensity within the area of interest. Focused beam conditions have previously been used to introduce distribution of displacement. For example, a framework performed in the HVEM-Tandem facility at Argonne National Laboratory in the US, employed dual beam irradiations with ions and electrons in an effort to observe the retardation effect of concurrent electron irradiation on ion-induced amorphization in Si.16) They found that the same intensity could be expected for an ion beam within an area of interest, while an electron beam could be focused. In their experiments, they also found that ion-induced amorphization was retarded by simultaneous electron irradi+1
Corresponding author, E-mail:
[email protected] Present address: Hitachi Research Laboratory, Hitachi 319-1292, Japan
+2
ation. The area close to the beam center, which was identical to the area irradiated with the high intensity electron beam, maintained crystallinity, while the outer area was amorphized. The amorphous/crystal interface is defined as the region where the amorphous cluster formation (amorphization) is balanced out by the epitaxial growth of the crystalline phase (recrystallization).17) This work was extended by Abe et al.18) to include quantification of the effect of displacement cascades on ion-induced amorphization at room temperature. Experiments of this type are valid, provided that the diffusion of point defects can be neglected. The electron beam intensity gradient in a focused electron beam is attributable to its distribution of displacements. Observations have shown that, at high temperatures, the diffusion of point defects due to the defect concentration distribution is especially prevalent at the off-center positions of the electron beam. This phenomenon was previously observed in a study that reported thickness contrast changes at the periphery of the beam.14,18) Simultaneous observation of crystalline-to-amorphous transformation induced by electron irradiation in semiconductors has been carried out as described in Ref. 19). The non-equilibrium new compound phase or characteristic two-phase structures formed were probably influenced by the high surface energy of the nanometer-scale particles. Carbides and nitrides consisting of high-Z elements,20,21) except for SiC,22) were found to be stable under irradiation. The geometries of those systems were characteristic, especially in the cases of nanometer-scale particles or thin films. On the other hand, the nanometer-scale particles, such as carbides or oxides embedded in steels, work as one of the factors enhancing material strength. In this system, we should concern about radiation effects in the particles and their interactions with the matrix, which are also influenced by radiation. In fact, one of the structural materials used in fusion reactors, F82H, contains MX (Ta or W for M, and C or
424
H. Abe et al.
N for X)-type precipitates whose radiation-induced instabilities have recently been reported.23) Oxygen dispersed strengthened steel (ODS steel) contains oxide particles, such as Y2O3, Y2TiO7, or other type of oxides, that are instable under irradiation as well.24,25) However, it should be noted that the particles are basically stable when they are irradiated independently. Therefore, it is reasonable to assume that displacement within a matrix influences the stability of the nanometer-scale particles. In this work, the diffusion of matrix vacancies is enhanced both by irradiation and diffusion based on the vacancy concentration gradient, which is achieved uniquely under focused electron irradiation in HVEM. 2.
Kinetic Equations of Point Defects
Dispersed particles and precipitates, hereinafter referred to as particles, in a metallic matrix such as the MX-type precipitates in steels and the oxide particles in ODS steels, are expected to play an important role in strengthening the material. When used as structural materials in radiation environments, such as fission and fusion nuclear reactors, the effects of radiation, especially the displacive ones, are of particular concern. The possible behaviors under irradiation of the particles can be hypothesized as follows: displacements in the particle directly eject the constituent atoms from the particles, displacements in the matrix increase the concentration of point defects (especially vacancies), and enhance the solute diffusion of the constituent atoms of the particles, so-called radiation-enhanced diffusion.13) In addition to those effects, non-displacive radiation, which is described as an energy transfer below the displacement threshold to atoms neighboring the point defects, induces the exchange of a position of an atom and a point defect at the matrix/particle interface, which can be categorized as radiation-induced diffusion.9) In this work, we will focus on the effect of radiation-enhanced diffusion. Kinetic equations under irradiation are described as follows in Refs. 7,10,26): dCv ¼ ·º Zvi ðMi þ Mv ÞCi Cv dt Mv Csv ðCv C0v Þ þ Dv r2 Cv
ð1Þ
and dCi ¼ ·º Zvi ðMi þ Mv ÞCi Cv dt Mi Csi Ci þ Di r2 Ci ;
(TEM) thin foils, it is reasonable to assume that the incident electrons collide with atoms once in the sample thickness at most. Therefore, cross section of defect formation is constant over the volume of interest. In this study, we will vary the degree of focus to achieve beam intensity variations. Therefore, the term ·º is a function of location, namely ·ºðrÞ where r stands for the distance from the center of the electron beam. In the third term, Csi and Csv depend on the sample geometry. In this study, the regions where the sample thickness could reasonably be presumed to be constant were chosen. This allowed us to neglect the geometry effect on the term. The fourth term Dr2 C can be neglected if we can assume homogeneous defect formation over the volume of interest. The HVEM irradiation experiments16,811,24) were performed based on this assumption, and it is possible that equivalent conditions were achieved by employing defocused beams. According to the study on defect kinetics,10) the ratio Cv =Ci could be on the order of 106 or much larger at the equivalent state. The target of this study is nanometer-scale carbides and oxides distributed in an iron matrix, where constituent elements such as tantalum and yttrium migrate in a matrix associated with vacancies. Interstitial atoms, on the other hand, nucleate dislocation loops at relatively low temperature, while the majority is annihilated at the sink (such as the surface of TEM thin foils), or so due to the recombination of vacancy and interstitial atoms. Attention must be paid to the geometry of TEM specimens, especially in the cases of thin and thick foils.11) In this study, we employed relatively thin regions in order to simplify the irradiation behavior. This effect will be taken into account in the future. Based on the discussion above, we rewrite eq. (1) with r dCv ðrÞ ¼ ·ºðrÞ Zvi ðMi þ Mv ÞCi ðrÞCv ðrÞ dt Mv Csv ðCv ðrÞ C0v Þ þ Dv r2 Cv ðrÞ:
In this equation, the solution for terms 2 and 3 can be obtained by simulation with proper assumptions. Simple geometries, such as the one employed in this study, may allow them to be regarded as constant. In this study, qualitative estimation of the fourth term will be investigated by in-situ observations under irradiation conditions, as described in the following section. 3.
ð2Þ
where C and M are concentration and mobility, respectively. Subscripts i and v stand for interstitials and vacancies, respectively. Superscript 0 stands for thermal equilibrium. ·º represents the damage rate, which is described as the product of displacement cross-section and flux. Zvi is a recombination cross section of an interstitial atom and a vacancy. The annihilation of point defects to the sink, such as the surface, is described as the product of the mobility and the annihilation cross section at the sink, as denoted by Mv Csv and Mi Csi . The first term is the defect production under irradiation. Because of the geometry of transmission electron microscope
ð3Þ
Experimental Procedure
First, TEM samples were prepared for use in determining the stability of MX-type precipitates in F82H steel and oxide particles in ODS steel. The instabilities of the particles in both materials were recently reported.23,24) Initially, a Fe-TaC model alloy was fabricated via the arc melting method from pure iron (99.997% in purity) and 0.2 mass% TaC powder (99.627%). The samples were then solution annealed at 1523 K for 24 h, and heat-treated at 1123 K for 1 h to precipitate TaC. Conditions were optimized to simulate the morphology of TaC particles in F82H steel. The final TEM samples were then produced by electrochemical polishing in 5% perchloric acid and 95% acetic acid at 288 K and 50 V.
New Approach to In Situ Observation Experiments under Irradiation in HVEM
The 9Cr-ODS steel with the designated composition of Fe 9Cr2W0.2Ti0.35Y2O3 (mass%) was supplied from National Institute of Fusion Science, Japan. The electrochemical polishing was done at room temperature in the same solution described above. Identification of the oxide particles was done by selected area electron diffraction and scanning transmission electron microscopy combined with energy dispersive X-ray spectroscopy (STEM/EDS) analysis and the particles, which were distributed inside of the grains, were identified as YTiO complex. Electron irradiation experiments were performed in Ultra High Voltage Electron Microscope (UHVEM) at Osaka University. The specimens were observed simultaneously undergoing irradiation with electrons at energy ranges from 0.75 to 2.5 MeV at 673 K. The bright-field (BF) image and selected area electron diffraction (SAED) were recorded either on slow-scan charge-coupled device (CCD) or imaging plate. The temperature was chosen to approximate the environment of the first wall material used in a fusion reactor, and in consideration of the reduced changes in the iron matrix that could be expected because the temperature tends to retard the evolution of dislocations, dislocation loops and voids. Electron flux at the center of the beam was taken as representative for the beam intensity throughout the experiment. Beam intensity variations were produced by adjustments to the degree of beam focus, which was fixed during the irradiation process. The potential for unintended 7
e⋅m s
-2 -1
6 5
Flux, φ / 10
23
4 3 2 1 0 -1000 -500 0 500 1000 Distance from the center of the beam, r /nm
Fig. 1 Intensity profile of a focused beam of 2.5 MeV electrons.
irradiation application to the area of interest during photo work and/or other beam adjustment activity was minimized by keeping that area away from the beam to the greatest extent possible. The intensity of the electron beam and its profile were measured using a Faraday cup located at the level of the viewing screen. Figure 1 shows typical intensity profiles of a focused electron beam. The intensity was reasonably approximated by a Gaussian distribution. 4.
Results and Discussion
Figure 2 shows a typical example of the microstructural evolution under electron irradiation in Fe-TaC and ODS steels. Decrease in size and diminishing in contrast were evident. In addition to those observations, a weakening in contrast of the particles and annihilation, as well as accumulation of voids around the oxides, were also observed in ODS. Change in particle size under irradiation with 2 MeV electrons is shown in Fig. 3. Particles were found to be reduced both in size and number density. At the early stage of irradiation, an especially notable linear shrinkage in size was achieved. Hereinafter, we define the size change rate as the change in size at the unit fluence: ðr2 r1 Þ=ðºt2 ºt1 Þ dr in unit of nm=ð1022 e=cm2 Þ; ¼ dðºtÞ
ð4Þ
where r1 and r2 are diameters of the particle of interest at the fluence of ºt1 and ºt2 , respectively (ºt2 > ºt1 ). Note that eq. (4) is the rate-per-unit fluence. If we neglect the diffusion of vacancies due to the vacancy concentration gradient, the amount of vacancies attributable to particle annihilation is a function of displacement in the vicinity of the particles. It turns out that the size change rate is independent of the electron flux. The size change rates under electron irradiation were measured under various beam intensities and energy levels. The particle locations were recorded and summarized as distances from the center of the electron beam. Figure 4 shows an example of the distribution of TaC particles. In this
(a)
0
4.1
9.7
21 x 1026 e/m2
0
1.6
4.8
8.8 x 1026 e/m2
(b)
Fig. 2
425
Microstructural evolution in (a) Fe-TaC and (b) 9Cr-ODS irradiated with 2 MeV electrons at 673 K.
426
H. Abe et al.
3
35
-26
dr/d(φ t ) / nm 10
Size, r /nm
25 20 15 10 5 0
48 X 1023 e /m2 s 26 X 1023 e /m2 s 15 X 1023 e /m2 s 10 X 1023 e /m2 s
2
-1
e m
30
2
B−1 B−2 B−3 B−4
0
5
10
15
Fluence, φ t /10
26
20
em
-2
25
1 0 -1 -2 -3
30
0
10
20
30
Flux, φ /10
23
Fig. 3 Size variation of TaC particles in Fe-TaC irradiated with 2 MeV electrons at 673 K. Four of the particles are shown as examples. Note that a roughly linear decrease in size was observed.
40 -2
em s
50
-1
Fig. 5 Flux dependence of the size change rate under irradiation with 2 MeV electrons. Four sets of beam intensity at the beam center are shown.
(a)
(b)
Y
Y
X
X
Fig. 4 Distribution of TaC particles in an iron matrix under irradiation with (a) 1.5 and (b) 1.0 MeV electrons on CCD image. The solid squares stand for the location of individual particles based on the system of coordinates with their origin at the center of the electron beam (arbitrary unit). The grayscale shows the size change rate. Particles with higher shrinkage rates are indicated with brighter contrast as well as arrows.
figure, the size change rate is shown in grayscale. Here, it can be clearly seen that a higher shrinkage rate was achieved in the vicinity of electron beam. Figure 5 shows the size change rate as a function of electron flux. The horizontal axis was determined by the combination of the beam intensity profile, as shown in Fig. 1 and the location of the particles. In other words, it is the local electron flux for the each TaC particles. As can be seen in the figure, the size change rate was almost constant when irradiated with relatively high intensity electrons, i.e., near the beam center. In contrast, flux dependence was especially evident at lower flux regime, i.e., in the vicinity of the beam. In Fig. 5, it was observed that the high intensity, i.e., focused beam, denoted with circles depends strongly on the flux. This is because the focused beam leaves a relatively steep distribution of vacancy concentration, which enhances vacancy diffusion comparing with a poorly focused beam. The energy dependence will be
described in detail elsewhere.27) Briefly, the above trend was found to be common under irradiation with electrons with energy levels above 1 MeV. The shrinkage rate became much higher under electron irradiation at energy levels higher than 1.5 MeV, while it was hardly influenced by electron irradiation at energy levels below 1 MeV, such as 0.75 and 0.2 MeV. The above results indicate the mechanism of irradiation integrity of TaC in an iron matrix where the gradient of electron beam intensity was achieved. Electrons higher than 1 MeV are attributable to atomic displacements in iron.8) It turns out that the atomic displacement vacancy concentration gradient results in a diffusion of vacancies from the beam center to the outer area. Therefore, the concentration of vacancies became relatively higher in the vicinity of the electron beam. The frequency of vacancy-tantalum complex formation increases, which enhances the vacancy-type
New Approach to In Situ Observation Experiments under Irradiation in HVEM
diffusion of Ta into an iron matrix. This is consistent with the finding that the high intensity beam was attributable to the higher shrinkage rate in the vicinity of the electron beam. The electrons with energy below 1 MeV hardly displace iron atoms, as no evidence of the diffusion effect was observed. On the other hand, the displacement of carbon in TaC is presumed less effective to the shrinkage of the TaC particles, because the non-stoichiometry, TaC1¹x, is allowed in the system. The weakening of oxide particles and/or the evolution of voids surrounding the oxides in the ODS steel, as shown in Fig. 2(b), are presumably based on the same mechanism, in which vacancies diffuse based on the gradient of their concentration. Some vacancies migrate to the oxide where they react with yttrium or titanium and assist them in dissolving into the matrix. Since the deviation from the stoichiometric composition is allowed in the system, the averaged mass deceases according to the accumulation of vacancies at the vicinity of the oxide particles. As a result, the oxides lose their Z-contrast and became indistinguishable. The formation of voids around the oxides provides evidence of the flow of vacancies towards the oxide particles. Once voids are formed, the change in oxides becomes relatively low because of the limited diffusion paths for the constituent atoms. The mechanism is discussed in detail in Ref. 28). 5.
Conclusions
This paper explored a new method of simultaneous observing specimens undergoing electron irradiation in HVEM in order to better understand the mechanisms related to the stability of nanometer-scale particles. In addition to the displacement of their constituent atoms, it was found that the intended distribution of vacancies in a matrix induced by focused beam irradiation results in enhancing diffusion of vacancies and possibly solute atoms from particles into the matrix. The instability of nanometer-scale particles such as TaC and Y2TiO7 were successfully observed while they were undergoing irradiation. Acknowledgments The authors are grateful to Profs. T. Muroga and T. Nagasaka, and Drs. H. Tanigawa and H. Sakasegawa for supplying the ODS and F82H steels used in our experiments. A portion of this work was supported by a project assigned to the Japanese Implementing Agency within the “Broader Approach Agreement”, by “R&D of nuclear fuel cladding materials and their environmental degradations for the
427
development of safety standards” by MEXT, and by the “Study on hydrogenation and radiation effects in advanced nuclear fuel cladding materials” under the Strategic Promotion Program for Basic Nuclear Research. REFERENCES 1) N. Yoshida and M. Kiritani: J. Phys. Soc. Jpn. 35 (1973) 14181429. 2) G. J. C. Carpenter and J. F. Watters: J. Nucl. Mater. 101 (1981) 2837. 3) S. Watanabe, N. Sakaguchi, N. Hashimoto and H. Takahashi: J. Nucl. Mater. 224 (1995) 158168. 4) H. Watanabe, T. Muroga, N. Yoshida and K. Kitajima: J. Nucl. Mater. 158 (1988) 179187. 5) D. Pêcheur, F. Lefebvre, A. T. Motta, C. Lemaignan and D. Charquet: J. Nucl. Mater. 205 (1993) 445451. 6) C. Kinoshita, H. Abe, K. Fukumoto, K. Nakai and K. Shinohara: Ultramicroscopy 39 (1991) 205212. 7) A. C. Damask and G. J. Dienes: Point Defects in Metals, (New York: Gordon and Breach, 1963). 8) N. Yoshida, M. Kiritani and E. F. Fujita: J. Phys. Soc. Jpn. 39 (1975) 170179. 9) M. Kiritani: J. Phys. Soc. Jpn. 40 (1976) 10351042. 10) M. Kiritani, N. Yoshida, H. Takata and Y. Maehara: J. Phys. Soc. Jpn. 38 (1975) 16771686. 11) M. Kiritani: Ultramicroscopy 39 (1991) 135159. 12) L. K. Mansur: Nucl. Technol. 40 (1978) 534. 13) L. E. Rehn and P. R. Okamoto: Nucl. Instrum. Meth. Phys. Res. B 39 (1989) 104113. 14) H. Abe, H. Naramoto and C. Kinoshita: Mat. Res. Soc. Symp. Proc. on Microstructure of Irradiated Materials 373 (1994) 383388. 15) H. Abe, N. Sekimura and T. Tadokoro: Mater. Trans. 46 (2005) 433 439. 16) D. N. Seidman, R. S. Averback, P. R. Okamoto and A. C. Baily: Phys. Rev. Lett. 58 (1987) 900903. 17) R. G. Elliman, J. S. Williams, W. L. Brown, A. Leiberich, D. M. Maher and R. V. Knoell: Nucl. Instrum. Meth. Phys. Res. 1920 (1987) 435 442. 18) H. Abe, C. Kinoshita, P. R. Okamoto and L. E. Rehn: J. Nucl. Mater. 212215 (1994) 298302. 19) H. Yasuda and H. Mori: Mater. Trans. 45 (2004) 24. 20) K. E. Sickafus, L. Minervini, R. W. Grimes, J. A. Valdez, M. Ishimaru, F. Li, K. J. McClellan and T. Hartmann: Science 289 (2000) 748751. 21) H. Abe: Thesis, (Kyushu University, 1993). 22) H. Inui, H. Mori and H. Fujita: Philos. Mag. B 61 (1990) 107124. 23) H. Tanigawa, H. Sakasegawa, H. Ogiwara, H. Kishimoto and A. Kohyama: J. Nucl. Mater. 367370 (2007) 132136. 24) I. Monnet, P. Dubuisson, Y. Serruys, M. O. Ruault, O. Kaitasov and B. Jouffrey: J. Nucl. Mater. 335 (2004) 311321. 25) I. Monnet, T. Van den Berghe and Ph. Dubuisson: J. Nucl. Mater. 424 (2012) 204209. 26) S. Ishino: Shousya Sonsho, (Univ. Tokyo Press, 1979) (in Japanese) pp. 148162. 27) H. Abe, T. Ishizaki, S. Kano, F. Li, Y. Satoh, T. Matsunaga, H. Tanigawa, D. Hamaguchi, T. Nagase and H. Yasuda: J. Nucl. Mater., submitted. 28) F. Li, H. Abe, T. Ishizaki, Y. F. Li, T. Nagasaka, T. Muroga, T. Nagase and H. Yasuda: J. Nucl. Mater., submitted.
