Accepted Manuscript Atomic structure of nano voids in irradiated 3C-SiC Y.R. Lin, L.G. Chen, C.Y. Hsieh, A. Hu, S.C. Lo, F.R. Chen, J.J. Kai PII:
S0022-3115(17)31054-1
DOI:
10.1016/j.jnucmat.2017.09.043
Reference:
NUMA 50530
To appear in:
Journal of Nuclear Materials
Received Date: 24 July 2017 Revised Date:
23 September 2017
Accepted Date: 28 September 2017
Please cite this article as: Y.R. Lin, L.G. Chen, C.Y. Hsieh, A. Hu, S.C. Lo, F.R. Chen, J.J. Kai, Atomic structure of nano voids in irradiated 3C-SiC, Journal of Nuclear Materials (2017), doi: 10.1016/ j.jnucmat.2017.09.043. This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
ACCEPTED MANUSCRIPT
Atomic Structure of Nano Voids in Irradiated 3C-SiC Y.R. Lin1, L.G. Chen1, C.Y. Hsieh2, A. Hu3, S.C. Lo2, F.R. Chen1 and J.J. Kai1,3,a) National Tsing-Hua University, Department of Engineering and System Science, 30013 Hsinchu, Taiwan
2.
Industrial Technology Research Institute, Material and Chemical Research Laboratories, 31040 Hsinchu, Taiwan
3.
The City University of Hong Kong, Department of Mechanical and Biomedical Engineering, 852 Kowloon, Hong
RI PT
1.
SC
Kong Special Administrative Region
Abstract
M AN U
It is important to understand the atomic structure of defect clusters in SiC, a promising material for nuclear application. In this study, we have directly observed and identified nano voids in ion irradiated 3C-SiC at 800 oC, 20 dpa through ABF and HAADF STEM images. A quantitative method was used to analyze HAADF images in which atomic columns with a difference in the
TE D
number of atoms could be identified and scattered intensities can be computed. Our result shows that these voids are composed of atomic vacancies in an octahedral arrangement. The density of
Highlights
EP
the void was measured by STEM to be 9.2x1019m-3 and the size was ~1.5 nm.
AC C
•We first directly observed octahedral voids in irradiated SiC at 800 oC. •A quantitative method was used to analyze number of atoms along an atomic column. •This observation provides better understanding of the evolution of microstructures and the mobility of vacancies in irradiated SiC. ___________________________
1. Author to whom correspondence should be addressed. Electronic mail:
[email protected]
1
ACCEPTED MANUSCRIPT
1. Introduction SiC is one of the most promising structural materials for next generation nuclear reactors.1, 2 It has been studied for a long time in the field of fusion reactor environments due to its low
RI PT
activation characteristics and good mechanical behavior at very high temperatures. After the Fukushima accident, SiC was identified as a primary candidate material for accident-tolerant nuclear fuel cladding.3 When SiC is irradiated at temperatures above 1000oC, vacancies tend to
SC
aggregate to become voids or intrinsic stacking fault loops, which are identified by transmission
decreases with elevated temperatures.6, 7
M AN U
electron microscopy (TEM). 4,5 The size of these voids increases and the number density
However, voids have not been observed in irradiated SiC under 1000 oC. In order to form a void, the temperature must be high enough for vacancies to migrate. In the case of SiC, Si vacancies
TE D
were shown to be mobile only when the temperature reached 750 oC.8 On the other hand, C vacancies were expected to be mobile even at higher a temperature.9 Studies have suggested that vacancies remain as a single point defect or di-vacancy defects inside the irradiated samples that
EP
are too small to be observed under TEM.10 Although atomic models of these irradiation-induced vacancy defects have been predicted, there is no direct evidence that confirms the atomic
AC C
configuration of these defects.
Recently, Cs-corrected STEM (scanning transmission electron microscopy) with ultra-high resolution was used to investigate those previously undefined defect in irradiated materials. When applied in a restricted zone-axis orientation, the high angle annular dark field (HAADF) scattering signal from a single column of atoms is strongly dependent on the atomic number (Zn, where n≈2), and the thickness of the sample. Hence, it is also called a Z-contrast image.11,12 The
2
ACCEPTED MANUSCRIPT
annular bright-field (ABF) imaging technique that collects lower-angle signals is able to directly detect the atom position of light atoms, such as carbon, which cannot be significantly imaged by
RI PT
HAADF images. 5,13,14
2. Experimental 2.1 Materials and irradiation conditions
SC
The material used in this study is a single crystal 3C-SiC grown on a Si substrate via chemical vapor deposition (NOVASiC, France). Specimens were irradiated with 5.1 MeV Si2+ ions using the DuET facility at Kyoto University, Japan. The irradiation conditions were set with a fluence
M AN U
of 5.65×1017 ion/cm2 at an irradiation temperature of 800oC under vacuum (10-6 torr), with the rate of displacement per atom at about 2.7 × 10-3 dpa/s at the depth regime of 20 dpa. The profile of the implanted Silicon ion was simulated with the SRIM program and the displacement energies of Si and C were 35 eV and 21 eV, respectively.6 In order to avoid the accumulation of the implanted Si ion in the SiC layer, we investigated the specimen’s microstructure at the depth
2.2 STEM technique
TE D
of 0.6 µm from the irradiated surface.
EP
In this study, we systematically characterized the atomic structure of the nano voids in an irradiated 3C-SiC using Cs-corrected STEM (JEOL, JEM-ARM200F) at an accelerating voltage
AC C
of 200 kV with ultra-high spatial resolution of ∼ 0.12 nm. Thin foil specimens for TEM analysis were prepared for cross-sectional observation by mechanical polishing followed by the 3-5 KeV Ar ion-milling method. Structural models of the atomic arrangement were constructed using CrystalKitX software.
Further, a recently developed method enabled us to measure the total intensity of the scattered electrons for each atomic column.12,15,16,17 In this study, HAADF image simulation was carried
3
ACCEPTED MANUSCRIPT
out with the commercial software MacTempas.12 An overview of parameters set for the image simulation is given in Table. 1. Here we constructed SiC structural models, viewed along the [011] direction, with different thicknesses (number of atoms). The intensity at each Si and C
RI PT
atomic-column in the simulated and experimental images was averaged within a circular mask, and the background intensity was also removed. The averaged intensities were recorded and demonstrated more clearly in Fig. 2 where the normalized intensity values are presented as a
SC
function of the number of Si and C atoms in a column together with a straight line through these values using linear regression. To extract quantitative information from the HAADF STEM
M AN U
image, certain parameters need to be kept constant for calibration purposes. The following equation calculates the ratio of the scattered intensities of the elements or compounds: I = I + × {[( / × ) × × ]} (1)
Where I is the intensity of the background, is the density of the material, is the atomic
TE D
weight, is the neutron mass, is the thickness of the sample, and are parameters, and is the atomic number. In addition, the Z-power law dependence of the intensity for two types of columns identified in the sample on their atomic number Z can be written as
∝ ( ) (2)
EP
AC C
According to equation (2), the Z dependent power law measured for this case is 2.12 when the sample thickness is 60nm.
3. Results
The HAADF image (Fig. 1a) and ABF image (Fig. 2a) were recorded simultaneously from the same area. In this area, abnormal and normal contrast regions viewed along the [011] zone-axis are indicated in different marker colors (Fig. 1b and Fig. 2b). The intensity of each bright spot
4
ACCEPTED MANUSCRIPT
was counted through ~20 peaks. For instance, in Fig. 1b, atom columns at the outer region of the rhombus shaped region were marked in red and the middle region were marked in blue. Lastly, the inner regions were marked in green. The total intensity of each peak was then depicted by a
RI PT
histogram of counts and intensities indicated in the color corresponding to each region.
Corresponding scattered intensities are presented in the histograms in Fig. 1c and Fig. 2c for HAADF images and ABF images, respectively. Because the thickness of the sample can be
SC
considered constant over the defective area, the absence of Si atoms and C atoms leads to a
decrease in intensity. The results show that the intensity of atom columns in both HAADF and
M AN U
ABF images has a decreasing and increasing gradient from the outer region to the inner region, respectively. Although ABF images provide a better image and precise position of the light carbon atoms in SiC, this imaging technique provided a partial Z contrast, such that the intensity of each atom column did not precisely reflect the number of absent atoms resulting from phase
TE D
contrast participating in the image.
Using the atom counting results shown in Fig. 3a, we can predict the number of atoms for each
EP
atom column and create a 3D reconstruction of this defect. According to the average intensity of atom columns in our HAADF experimental images (Fig. 1&3b) and the atom counting results,
AC C
the number of Si atoms and C atoms along a single atom column could be evaluated. Each single atomic column in Fig. 1a was analyzed using the same method. Knowing the number of atoms and presuming that the defect’s position is in the middle of the sample, the reconstructed structure shows that this defect should be an octahedral void missing 66 silicon atoms and 64 carbon atoms (Fig. 4a). When viewing along [011] (Fig. 4b), the defect with absent atoms colored in gray and rhomboidal in shape, bound by {111} planes, has a diagonal measuring ~1.3
5
ACCEPTED MANUSCRIPT
nm. As for the view along [011] (Fig. 4c), the defect in the rhomboidal shape has a diagonal measuring ~1.8 nm. Furthermore, the number density of the defects in the specimen was quantified based on the thickness (~60 nm) estimated by using an intensity of zero loss and
counted under experimental STEM images to be 9.2x1019 m-3.
SC
4. Discussion
RI PT
plasma loss in electron energy loss spectroscopy (EELS). The density of this nano-void was
In Fig. 3a, the EELS result shows that the sample thickness is about 60nm (195 atom). However,
M AN U
the normalized averaged intensity of an atomic column in the experimental images at this thickness is slightly lower than the simulated intensity. This may be due to the de-channeling effect which causes the loss of energy on the column. In this case, the number of atoms in a column and the column intensity will no longer maintain a linear relationship, and the curve flattens off instead. Furthermore, the position of the defect (near or away from the surface) and the strain caused by the void will also affect the intensity. Nevertheless, in this work both ABF
TE D
(Fig. 2) and HAADF (Fig. 1) images viewed along [011] zone-axis shows that the atomic columns’ intensity decrease symmetrically from the inner region (a rhombus region bounded by {111} planes) to the outer region. Therefore, the defect we observed should definitely be a vacancy type defect, which consist of vacancies bounded by {111} planes. It is worth noting that
EP
the octahedral cluster shape bounded by {111} planes is the most common experimentally observed geometry for vacancy aggregates present in the crystal grown silicon crystal18. In
AC C
addition, the size of formation energies for vacancy clusters that follow the octahedral motif has been calculated. The overall vacancy cluster energetics were described very well by a power law function that scales as Nv2/3, where Nv is the number of vacancies in the cluster.19 In other words, only if the vacancy cluster is small enough (<2nm) can octahedral voids take shape.
According to recent MD simulations, migration energies of a silicon vacancy, a silicon interstitial, a carbon vacancy, and a carbon interstitial are 2.35, 1.53, 4.10, and 0.74 eV,
6
ACCEPTED MANUSCRIPT
respectively.21 Therefore, the energies for diffusion of either the Si or C vacancy must be quite high. Based on the results of previous works, 8,9 at 800 oC, for vacancies in SiC, only Si vacancies were proved to be mobile. Further, it is generally accepted that vacancies can be produced either
RI PT
in single point defects or in clusters. Studies assumed that vacancy clusters surviving from cascades will soon be destroyed by self-interstitials, which is consistent with experimental
observations that dislocation loops are mainly of the interstitial type.22,23 However, we showed
SC
that nano-voids at a low density indeed exist in ion irradiated SiC at 800 oC. The density of the nano-voids (9.2x1019 m-3) is nearly three orders of magnitude smaller than the maximum density
M AN U
of voids observed above 1300 oC in previous works.6,7 The lower density of nano-voides suggests that cavity nucleation is relatively rare at 800 oC due to the limited mobility C vacancies.
5. Conclusion
TE D
In this study, octahedral nano voids in ion irradiated 3C-SiC at 800 oC were first observed and characterized through the Cs-corrected STEM. This microscope provided images of individual silicon and carbon atoms by the ABF method and HAADF images, which showed atomic
EP
number contrast for high scattering angles of the electrons. These characterization techniques further enhanced the 3D reconstruction of octahedral nano voids to gain a better understanding of
AC C
the microstructures of ion-irradiated single crystal SiC. In addition, we showed that voids can still be formed in irradiated SiC at temperatures where vacancies on only one sublattice are mobile.
Acknowledgements This work was supported by the Bureau of Energy, Ministry of Economic Affairs (Taiwan) Program NSC104-AB-001-EJ. Electron Microscopy was performed with the JEM-ARM200F
7
ACCEPTED MANUSCRIPT
microscope at the Material and Chemical Research Laboratories, which is supported by the Industrial Technology Research Institute, Taiwan. We thank Tsai-Tian Wu for his assistance
RI PT
with sample preparation.
References 1.
Y. Katoh, L. L. Snead, I. Szlufarska, and W. J. Weber, Curr. Opin. Solid St. M. 16, 143
L.L. Snead, T. Nozawa, M. Ferraris, Y. Katoh, R. Shinavski, and M. Sawan, J. Nucl. Mater. 417, 330 (2011).
3.
M AN U
2.
SC
(2012).
Y. Katoh, L. L. Snead, C. H. Henager, T. Nozawa, T. Hinoki, A. Iveković, S. Novak, and S. G. Vicente, J. Nucl. Mater. 455, 387 (2014).
C. Y. Ho, S. T. Tsai, H.T. Lin, F. R. Chen, and J. J. Kai, J. Nucl. Mater. 443, 1 (2013)
5.
Y.R. Lin, C.Y. Ho, C.Y. Hsieh, M.T. Chang, S.C. Lo, F.R. Chen, and J.J. Kai, Appl. Phys.
TE D
4.
Lett. 104, 121909 (2014)
Y.R. Lin, C.K. Ku, C.Y. Ho, W.T. Chuang, S. Kondo, and J.J. Kai, J. Nucl. Mater. 459, 276 (2015).
EP
6.
S. Kondo, Y. Katoh, and L.L. Snead, J. Nucl. Mater. 386, 222 (2009).
8.
H. Itoh, N. Hayakawa, I. Nashiyama, E.Sakuma, J. Appl. Phys. 66, 4529–4531 (1989).
9. 10.
AC C
7.
L. Snead & Y. Katoh, Comprehensive Nucl. Mat. 215-239 (2012)
T. Ohshima, A. Uedono, K. Abe, H. Itoh, Y. Aoki, M. Yoshikawa, S. Tanigawa, and I. Nashiyama, Appl. Phys. A 67, 407 (1998)
11.
Y.R. Lin, C.Y. Ho, W.T. Chuang, C.K. Ku, and J.J. Kai, J. Nucl. Mater. 455, 292 (2014).
12.
F.R. Chen, D. Van Dyck, and C. Kisielowski, Nat. Commun. 7, 10603 (2016)
8
ACCEPTED MANUSCRIPT
13.
P. Hartel, H. Rose, C. Dinges, Ultramicroscopy 63 ( 1996) 93- 114
14.
P. J. Phillips and R. F. Klie, Appl. Phys. Lett. 103, 033119 (2013)
15.
S.D. Findlay, N. Shibata, H. Sawada, E. Okunishi, Y. Kondo and Y. Ikuhara,
RI PT
Ultramicroscopy 110 (2010) 903-923 16.
L. Jones, Mater. Sci. Eng. 109, 012008 (2016).
17.
H. Fukuda, M. Yasuda, and T. Iwabuchi, Appl. Phys. Lett. 61, 693 (1992)
18.
L. Yedra, P. Torruella, A. Eljarrat, A. D. Darbal, J. K. Weiss, F. Peiró, and S. Estradé,
SC
Appl. Phys. Lett. 105, 053117 (2014)
T. Sinno, E. Dornberger, W. von Ammon, and F. Dupret, Sci. Eng. R 28, 149-198 (2000)
20.
M. Prasad and T. Sinno, Appl. Phys. Lett. 80, 1951 (2002)
21.
F. Gao, W. J. Weber, M. Posselt, and V. Belko, Phys. Rev. B 69, 245205 (2004)
22.
M. Ishimaru, Y. Zhang, S. Shannon, and W. J. Weber, Appl. Phys. Lett. 103, 033104
M AN U
19.
(2013)
H. Huang and N. Ghoniem, J. Nucl. Mater. 250, 192 (1997).
TE D
23.
EP
Table Captions
AC C
TABLE I. Overview of parameters set for the image simulation
Voltage
200 KeV
Probe size
1.2 Å
Probe Semi-Angle
27 mrad
Inner Aperture
60 mrad
Outer Aperture
160 mrad
9
ACCEPTED MANUSCRIPT
M AN U
SC
RI PT
Figure Captions
TE D
Figure 1: HAADF-STEM image of an octahedral nano void in 3C-SiC, viewed along the [011] zone axis. (a) HAADF image. (b) Atom columns marked in specific color. (c) Histogram of
AC C
EP
scattered intensities of atom columns marked in specific color of Figure 1(b).
10
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
Figure 2: ABF-STEM image of an octahedral nano void in 3C-SiC, viewed along the [011] zone axis. (a) ABF image. (b) Atom columns marked in specific color. (c) Histogram of scattered
AC C
EP
TE D
intensities of atom columns marked in specific color of Figure 2(b).
11
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
Figure 3: Quantification of HAADF-STEM images of 3C SiC. (a) Linear increase of the
TE D
estimated mean intensity values with increasing number of Si (blue) and C (yellow) atoms in a column oriented along [011]. (b) Intensity of atom columns in HAADF experimental images
AC C
EP
with sample thickness ~60 nm (195 atoms) oriented along [011].
12
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
TE D
Figure 4: Schematic of an octahedral nano void. (a) In the 60nm thickness sample. (b) In the view of [011] zone-axis orientation. (c) In the view of [001] zone-axis orientation. The bigger
AC C
EP
gray ball stands for the Si vacancy and the smaller for the C vacancy.
13