Materials Transactions, Vol. 46, No. 3 (2005) pp. 433 to 439
Special Issue on Fusion Blanket Structural Materials R&D in Japan
#2005 The Japan Institute of Metals
Stability and Mobility of Interstitial-Type Defect Clusters
Generated from Displacement Cascades in Copper and Gold
by In-Situ Transmission Electron Microscopy
Hiroaki Abe1;2 * , Naoto Sekimura2 and Tadayasu Tadokoro2
1
2
Research Center for Nuclear Science and Technology, University of Tokyo, Tokai, Ibaraki 319-1188, Japan
Department of Quantum Engineering and Systems Science, University of Tokyo, Tokyo 113-8656, Japan
In-situ TEM observations were carried out in copper and gold under 100 keV Cþ , 240 keV Cuþ , 600 keV Kr2þ , and 900 keV Xe3þ
irradiations from 573 to 823 K, in order to obtain direct experimental insights into the defect accumulation processes. Defect clusters
corresponding to displacement cascades were observed to be unstable depending on the temperature, ion species, and fluence. Multiple (2 or 3)
defect clusters showing up with their contrast in the same video frames were considered to be features associated with subcascades and high
mobility of these clusters when located within 20 nm and from 20 to 140 nm, respectively. Instability and diffusion of defect clusters were also
detected under ion irradiation. The directions of the cascade-driven and instability-driven motions of the defect clusters were strongly related to
crowdion directions, suggesting that the underlying mechanism is based on the motion of crowdion-related glissile defects. This instability is
interpreted as the transition of sessile defects into glissile ones. The effects of intra- and inter-cascades on the formation of glissile defects are
suggested based on their dependence on the ion species and flux.
(Received September 27, 2004; Accepted December 15, 2004)
Keywords: transmission electron microscopy, in-situ observation, ion irradiation, displacement cascades, copper, gold
1.
Introduction
Energetic particles, such as ions and neutrons, transfer a
portion of their energy to lattice atoms in the form of kinetic
energy; this leads to lattice displacements. When kinetic
energy of more than a few keV is transferred to a primary
knock-on atom, dense displacements, so-called a displacement cascade is generated from a local volume of the order of
103 to 104 atoms. According to molecular dynamic (MD)
simulations,1,2) a volume can be locally and spontaneously
melted and rapidly cooled to leave vacancies and interstitial
atoms. The defect clusters may therefore be formed after the
cooling phase and can be simultaneously observed with a
transmission electron microscope interfaced with an ion
accelerator (TEM-Accelerator).3–5) This method is referred to
as the indispensable method. The simultaneous observations
of the accumulation process and the instability of defect
clusters under ion irradiations revealed that the fraction and
size of surviving defects depend on irradiation conditions,
such as incoming particles and temperature.6–13) In-situ
observations using high-voltage electron microscopes are
also indispensable in detecting the motion of defect clusters
in metals.14,15) The motion suggests that dislocation loops
transform into mobile defects enabling diffusion with
relatively low activation energy; however, its mechanism
has not yet been experimentally clarified. Recently conducted
MD simulations have revealed the generation of glissile
defects consisting of crowdions in displacement cascades in
Cu and Fe.16,17) The glissile defects are capable of traveling
large distances in the crowdion directions, i.e., h110i in fcc
and h111i in bcc. The activation energy is estimated to be
extremely small, below 0.04 eV, which is even smaller than
the binding energy of the defects and the migration energy of
single interstitials.18) The MD simulation results indicate that
*Corresponding
author, E-mail: [email protected]
the crowdion-related interstitial-type defect clusters can be
formed within displacement cascades, and they are able to
diffuse rapidly but are unstable especially at relatively high
temperatures. Evidently, there is no report that verifies the
existence of interstitial-type defects in copper and gold at
temperatures above 500 K; however, our recent in-situ TEM
observations have provided evidence for these defects even at
temperatures of 700 K and above.19) However, the characteristics of the glissile defects and related cascade effects are
still unclear.
In this work, we present simultaneous microscopic
observations under ion irradiations using the TEM-Accelerator facility, in order to reveal the formation, stability, and
nature of mobile (glissile) defect clusters as well as to obtain
an insight into inter- or intra-subcascade effects on the
glissile defect clusters.
2.
Experimental Procedures
Pure copper (>99:999%) and pure gold (99.99%) were
separately annealed at 1000 K for 12 h in Ar-3% H2 gas flow
and in air, respectively; they were then cut into disks of 3 mm
in diameter. Subsequently, to achieve electron-transparent
thin foils, the Cu and Au specimens were electrochemically
perforated in a solution of 30% nitric acid in methanol at 20 V
and 277 K and in 1 mol LiCl in methanol at 30 V and 300 K,
respectively.20) Prior to the irradiation, samples were set in
TEM and annealed at 773 K or above to stabilize the surface
morphology during irradiation at high temperatures. This was
done because of the presence of a rather high vapor pressure
in vacuum and the irradiation-enhanced evaporation of the
samples. Samples whose surface normal was close to {011}
were taken to achieve stable surface morphologies,21) and in
several cases, in-situ observations in {001} samples were
carried out. The samples were then irradiated with ions in a
TEM (JEM-4000FX, JEOL Ltd.) with ion accelerators,
434
H. Abe, N. Sekimura and T. Tadokoro
Fig. 1 Microstructural evolution in copper irradiated with 240-keV Cuþ ions at 828 K. Ion beam current for the sample is 40 pA/
0:25 mm (5:1 1011 ions/cm2 s). The sequence is captured from videotapes with a time resolution of 1/30 s. The black arrows showing
two or three defects formed in the same time bin (video frame) appear relatively further apart.
installed at Takasaki Ion Accelerators for Radiation Application (TIARA), at the Japan Atomic Energy Research
Institute (JAERI), Takasaki, Japan.5,22) Irradiation in copper
was performed with either 100 keV Cþ ions and a current of
2:5 1013 C/cm2 s or 240 keV Cuþ ions and 5:1 1011 Cu/
cm2 s, while that in gold was performed at either 600 keV
Kr2þ ions and a current of 5:1 1010 Kr/cm2 s or 900 keV
Xe3þ and a current of 9:8 1010 Xe/cm2 s, at temperatures
ranging from 573 to 823 K. The beam transport was carefully
tuned using a mass analyzing magnet, a bending magnet, two
sets of electrostatic prisms, and an einzel lens with the aid of
a set of beam slits, a beam attenuator, a set of beam shifters,
and a set of beam scanners so as to achieve a well-aligned and
stable beam with a low-divergence at the sample position in
TEM. The beam area for the sample is roughly 1.5 mm in
diameter, and beyond this, the beam intensity is mostly
identical. The beam current is measured using a Faraday cup
with an entrance aperture diameter of 0.25 mm at the level of
the specimen. Considering the irradiation geometry, wherein
the angle between the ion beam and the sample normal is 30
degrees, the projected ranges from the sample surface were
estimated by TRIM (or SRIM) calculations.23) In copper,
these ranges corresponded to 115 and 57 nm for 100 keV Cþ
and 240 keV Cuþ ions, respectively. Since the thickness of
the observed regions is typically 50–70 nm, approximately
81–89% and 31–58% of the incoming Cþ and Cuþ ions
penetrate the sample, respectively. For Kr2þ and Xe3þ ions in
gold, the projected ranges are 87 and 85 nm, and the portions
of transmitted ions for a 50-nm-thick foil are 80% and 78%,
respectively. Microstructural evolution was observed at the
TEM operation voltage of 400 kV mainly by the weak-beam
dark-field (WBDF) technique at the conditions of n ¼ ½011,
g ¼ 200, g (4–6 g). The imaging system (Gatan Inc., model
676) consists of a YAG scintillator screen that is fiberoptically coupled to an image intensifier and further coupled
to a conventional TV-rate CCD (frame rate of 1/30 s). An
image of the sample, approximately equivalent to
108 nm 140 nm, was videotaped without image processors
so as to achieve the maximal time resolution. The effect of
remanence of the scintillator is not experimentally detected
in the system. The only two parameters that were processed
by a computer after image acquisition were contrast and the
brightness of each frame.
3.
Results and Discussion
3.1 Evolution of displacement cascades
The defect contrast features appeared and were observed
as dots or black-and-white contrast features. The diameter of
these defect contrast features ranged from 4 to 6 nm, as
shown in Fig. 1. The defects are presumed to be either
vacancy-type dislocation loops, stacking fault tetrahedra
(SFT), or interstitial-type dislocation loops. In copper, the
probability of defect cluster formation was estimated to be
2% and 67% of the incoming particles for Cþ and Cuþ ion
irradiations, respectively. In gold irradiated with Kr2þ and
Xe3þ , this probability was approximately 70%. In addition to
the single contrast feature, multiple (typically 2 or 3) contrast
features appeared in the video frames, each of which was
located within the distances range from 4 to 140 nm, as
shown in Figs. 1(b) and (f). Since the magnification of TEM
observations was fixed throughout the experiments, we could
not detect features that were located more than 150 nm apart,
which is approximately a diagonal of the measured area.
One characteristic feature of high-energy and heavy ion
irradiation is the branching off of the displacement sequences, forming multiple displacement-dense regions along
the trajectory of ions; these are referred to as subcascades.
The WBDF method employed in this work is a good measure
to observe the subcascades since it enhances the contrast
around the defects, resulting in a higher resolution for
distinguishing closely located contrast features.24,25) The
features located within a distance of 20 nm were regarded as
defects generated from the subcascades.19) Multiple contrast
Stability and Mobility of Interstitial-type Defect Clusters Generated from Displacement Cascades in Copper and Gold
features appeared in single video frames and were characteristic of heavy-ion irradiations, while no subcascades have
been detected from light ion irradiations such as carbon ions.
The probability that multiple independent ions pass through
and form defect clusters within such small areas is far smaller
than detected. Although the flux for C-irradiation was more
than double that of Cu and other heavy ions, this was notified
only in heavy ion irradiations. The distances between the
multiple defects agreed well with the estimation obtained
from TRIM.19,26)
On the other hand, as indicated with arrows in Fig. 1, a
large number of defects were evidently distributed over
distances exceeding those expected from subcascade events.
The phenomenon, hereafter referred to as multi-defect
formation, can be hypothesized in either of the following
ways: (i) the clusters are formed from independent ions,
hereafter referred to as multi- (or 2-) particle event. (ii) The
clusters are formed from an ion, and either or all clusters are
capable of traveling by a diffusion mechanism that is
currently undefined, hereafter referred to as single-particle
event. Figures 2(a) and (b)19) show the distribution of
distance between the defect clusters appearing in the same
video frame and the yield of directions, respectively. The
direction was defined as counterclockwise in the segment
between the two defects and as horizontal in the video frame.
The data in the figure were accumulated for a total irradiation
time of 10 min; however, due to a change in the TEM
contrast, the net accumulation time, in this case, was less than
5 min. It should be noted that the distance and direction are
Fig. 2 Distribution of the multiple-defect clusters that appeared in the
same time bins (the same video frames) as a function of distance in (a) and
angle in (b) in copper irradiated with 240-keV Cuþ ion flux of 5:1 1011
ions/cm2 s at 828 K.18) The defects appearing at distances below 20 nm in
(a) are presumably related to the subcascades. Anisotropic distribution is
seen in (b), and each peak corresponds to a projected h110i direction. Refer
to the text for a detailed discussion.
435
the ones projected on the imaging screen. The distance
distribution ranges up to 140 nm, and the yield of directions is
anisotropic, especially for features that encompass a diameter
of more than 20 nm. The frequency of a 2-particle event in a
region of, for example, 100 nm, is 0.80. However, the
possibility of the multi-particle event (i) can be eliminated for
the following reasons. One is that, as shown in Fig. 2(b), the
anisotropic distribution of multi-defect events is evident. The
three peaks in this figure are almost identical to the
projections of the h110i crowdion directions. It should be
noted that due to weak-beam conditions, the directions
measured in this work approximately deviated by 5 to 10
degrees from the identical 011 pole. The difference in the
yield at each h110i-related peak could be influenced by the
surface effects as well as by the reasons proposed by
experiments, such as the limitation of the analysis area and
the projection of three-dimensionally distributed clusters. In
the case of the (001) gold sample, we detected two peaks in
perpendicular directions along [110] and [11 0]. The other
reason is that the distance between such defects varies with
the irradiation time.19) At the beginning of irradiation, the
distance is up to 140 nm and it decreases with an increase in
the irradiation time. The fluence dependence of the formation
rate was also detected. Based on these reasons, we conclude
that the evolution of multi-defect formation in the same time
bin (video frame) results from a single particle event.
3.2 Instability of defect clusters
All the defects eventually disappeared within seconds at
the investigation temperature. Their lifetime depended on the
incident particles, temperature, and irradiation fluence.19) The
typical feature of the lifetime is as follows. Defects generated
from light ions disappeared in shorter lifetimes, typically
lasting for less than 3 s. On the contrary, defects generating
from heavy ones can have three lifetimes—less than 1 s, 3 to
several seconds, and longer lifetime—each of which is found
to be correlated with the nature of the defect clusters. A
typical lifetime of SFT was several seconds or more, and that
of the vacancy-type dislocation loops (V-clusters) was less
than several seconds, as estimated in a previous work.10) In
addition to the above, we must note that tiny defects (a couple
of nm) with much shorter lifetimes, lasting for several
thirtieths of a second or less, were detected. These have not
been previously reported, presumably because the WBDF
technique enables higher spatial and time resolutions than
those obtained by the bright-field technique. It should be
noted that many of the defects were associated with the
abovementioned multi-defect event. In Fig. 1, two defects,
100 nm apart, first appeared in (b) and then disappeared in 3/
30 or 4/30 second. Similarly, two of three defects appeared in
(f) and then disappeared in several thirtieths of a second.
The above results were derived from in-situ observations
under ion irradiations. To confirm the thermal stability of the
defects and to clarify the nature-dependent lifetime in gold,
we performed post-irradiation thermal annealing experiments
under the beam-off condition.8,10,11) This experiment considers in-situ observations as a function of time after ion
irradiation has been terminated. At the moment of beam off,
nonfatal mechanical vibrations were induced from an insert
of a Faraday cup for several thirtieths of a second. Image
436
H. Abe, N. Sekimura and T. Tadokoro
Cluster Number density, D/10 -14m-2
5
573 K
4
3
2
673 K
1
0
20
40
60
80
100
120
140
Time after beam off, t/s
Fig. 3 Change in the number density of defect clusters after ion irradiation.
Annealing curves for gold are shown at 573 K and 673 K.
drifting due to the beam-off did not interrupt the detailed
observation because of a very low beam current, in another
words, due to negligible beam heating. Figure 3 shows the
change in the defect area density at different temperatures as
a function of time, following the beam off condition. It is
known that there are two and three lifetime components at
573 K and 673 K, respectively, indicating the varying nature
of defect clusters. Ishino et al.10,27) claimed that vacancy-type
loops and SFT could be observed by a combination of the
bright-field and inside-outside techniques. We observed the
same behavior in ion-irradiated gold. In addition to these
defects, we observed defect clusters that spontaneously
disappeared in the beam-off condition because of their
extremely short lifetime, as shown in Fig. 3. By considering
the distinct difference in the lifetime and the irradiation
temperature (stage IV to V), we conclude that the short-
lifetime clusters are of the interstitial type (I-clusters).
Figure 4 shows the lifetime of defect clusters as a function
of irradiation temperature and ion species. Under Arþ
irradiation, the distribution of lifetime appears to extend
because the long lifetime defect clusters, namely V-clusters,
became more stable, while only a fraction of the shortlifetime ones, I-clusters, decreased at higher temperatures.
On the contrary, under Xe3þ irradiation, I-clusters became
less stable and the defects with a lifetime of less than 1 s
occupied more than 80% of the total number of defects.
Furthermore, the V-clusters occupy an extremely low
fraction. This difference can be described in terms of defect
distribution in a displacement cascade and diffusivity of point
defects. Suppose the threshold energy for subcascade
formation is 20 keV. The number and distribution of
subcascades in 100 nm-thick gold under 600 keV Ar2þ and
900 keV Xe3þ ion irradiation is estimated by TRIM. The
Ar2þ ions leave 2 subcascades that are, on an average, 13 nm
apart. Vacancies could agglomerate into V-clusters because
of the low defect density in rather extended volumes of
displacement cascades and higher yields in interstitial atoms
diffusing toward surface. Due to the sample thickness
employed in this work, the behavior of point defects should
be treated as a thin foil case. On the contrary, the Xe3þ ions
will leave 12 subcascades that are, on an average, 6 nm apart.
A high probability in I-cluster formation is presumed in dense
cascades; this is attributable to the formation of short lifetime
defects, as shown in Fig. 4. Among the densely packed
subcascades, intra-cascade reactions of interstitial defects,
e.g., coalescence, are introduced.28) This will be discussed in
detail later.
3.3 Mobility of defects (sessile-to-glissile transition)
Some of the stable (sessile) defects eventually get transformed into mobile (glissile) ones, as shown in Fig. 5. Based
600 keVAr2+
80
5.7 1014
60
673 K
Fraction (%)
40
723 K
773 K
ions/m2s
798 K
20
0
900 keVXe3+
80
9.7 10 14
ions/m2s
60
673 K
40
723 K
773 K
798 K
20
0
0
5 10 20 0
5 10 20 0
5 10 20 0
5 10 20
Lifetime of defect clusters, t/s
Fig. 4 Lifetime of defect clusters under ion irradiations at different temperatures. The data were taken during the first 30 s from the start of
irradiation.
Stability and Mobility of Interstitial-type Defect Clusters Generated from Displacement Cascades in Copper and Gold
437
Fig. 5 Mobility of glissile defects in gold under irradiation with 600 keV Ar2þ ion (1:3 1011 ions/cm2 s) at 723 K. The sequence is
captured from videotapes with a time resolution is 1/30 s. Onset of motion and rest position are shown with arrows in (b) and (h), and (c)
and (i), respectively. The short lifetime of defects formed through glissile-to-sessile transition is indicated with arrows in (g) and (k).
Moving distance, d/nm
120
(a)
(b)
100
80
60
40
20
0
0
30
60 90 120 150 180
0
(c)
10
30
60 90 120 150 180
(d)
8
Yield (arb)
on a discussion in a previous work,19) the Burger’s vectors of
sessile and glissile defects are presumed to be b ¼
1=3h111if111g and b ¼ 1=2h110if111g, respectively. We
could not apply the conventional g:b analysis and other
techniques for identification because of the instability of the
defects. The mobile defects were typically small (several nm
in diameter or smaller) and less stable (lifetime < several
thirtieths of a second), and observed in areas with a thickness
of 50-100 nm. Since the time-dependent formation rate
differed to a large extent from that of V-clusters, it is
suggested that these clusters are I-clusters, as discussed in the
previous sections.
The typical features of mobile defects in terms of mobility
were one-dimensional back-and-forth migration with an
amplitude of approximately 5 nm and a frequency of 1 s1
or more and one-directional motion with velocities ranging
from 50 to >103 nm/s. At the onset of motion, some of the
defects lose their strong strain contrast. Such a contrast
transition may result from the sessile-to-glissile transition.
However, it is noteworthy that a similar phenomenon can be
observed due to the persistence of vision in a frame (1/30 s) if
a high-contrast object moved fast enough. In addition to the
above features, mobile defects with features other than these
were observed. They are characterized by an approximate
size of 10 nm or more and defects with a staking fault
contrast. Since they appear to grow under ion irradiations,
which are indirectly related to displacement cascades, they
are omitted from this work.
The directions of their motion appeared to be anisotropic,
as shown in Fig. 6. The mobility of defects in this figure was
taken in the first few minutes under ion irradiation. The two
peaks observed in this figure correspond to the projections of
110 directions. Occasionally, three peaks of 110 projections
6
4
a
2
0
0
30
60 90 120 150 180
0
30
60 90 120 150 180
Direction (degrees)
Fig. 6 Distribution of moving distance in {011} gold irradiated with
600 keV Kr2þ ion flux of 5:1 1011 ions/cm2 s at temperatures of 723 K in
(a) and (c), and of 773 K in (b) and (d).
were observed probably due to the orientation relationship
with the surface, which works as strong sink for point defects.
In the case of the {001} sample, two peaks were observed
along 110 projections at an angle of 90 degrees. The same
phenomena were observed throughout the experiments in this
work. Figure 7 shows the time dependence of the defect
motion. It should be noted that a time bin is set for analysis. It
is evident that the moving distance decreases with an increase
in the irradiation fluence, and the distributions of Kr2þ and
438
H. Abe, N. Sekimura and T. Tadokoro
1.E-01
Frequency per ion
1.E-02
1.E-03
1.E-04
Fig. 7 Distribution of moving distance as a function of irradiation time in
gold irradiated with 600 keV Kr2þ and 300 keV Cuþ ions at 723 K.
1.E-05
1.E+14
1.E+15
1.E+16
1.E+17
Ion Flux, φ / ions cm-2s -1
Fig. 9 Frequency of sessile-to-glissile transition per ion as a function of
irradiation flux in gold irradiated with 300 keV Cuþ ions at 773 K.
Fig. 8 Frequency of sessile-to-glissile transition per ion as a function of
irradiation temperature and projectile.
Cuþ are identical. The formation rate of the mobile clusters
was observed to be dependent on fluence. The highest
formation rates coincided with 3 1014 Cu/cm2 and 1:5 1013 Kr/cm2 , respectively, and the underlying mechanism is
yet to be determined. We suspect that the size of defect
clusters plays an important role in their kinetics.
The formation rate of the mobile defects as a function of
irradiation temperature, ion species and ion flux is investigated as shown in Figs. 8 and 9. The high formation rate for
heavy ion irradiations and a monotonous decrease in the
formation within the investigated temperature can be clearly
seen in Fig. 8. This phenomenon can be described in terms of
subcascades, as employed in the discussion in Fig. 4. Simply
applying the TRIM, the probability of 20-keV subcascade
formation is 12 and 2 for Xe and Ar, respectively in the
investigated thickness. However, according to the experimental results as shown in Fig. 8, the ratio of the formation
probability of glissile clusters for Xe against Ar is approximately 2.5 to 4 depending on the temperatures. The result
strongly suggests elimination of the glissile clusters among
densely distributed subcascades within the investigated
conditions. The result of this model is consistent with the
MD simulation results wherein the bundled crowdions are
formed at the periphery of the displacement cascades.16) At
present, additional data is not available with us to discuss this
model; therefore, further investigation is required. Figure 9
shows the probability of sessile-to-glissile transition per ion
(p) plotted against the ion flux (). A slightly nonlinear
(p / 1:4 ) relationship in the case of Cuþ ion irradiation is
evident. The linearity is generally interpreted that a short
lifetime object transforms into the other by an impact, and the
deviation from it deduces reactions among individual objects
and impacts.29) The results indicate that the pre-existing
sessile clusters transform into glissile ones by displacement
cascades, and that the transition and the mobility is enhanced
by the inter-cascade reaction,29) presumably through energy
or momentum transfer. Another expected factor on the nonlinearity is the change in stability of the interstitial-type
defect clusters influenced by evolution of defects as shown in
Fig. 7.
A model for the observed phenomena is proposed as
follows.19) Glissile defects, presumed to be consisting of a
bundle of crowdions, are formed during the displacement
cascade process. Due to the low activation energy for the
crowdion bundle motion,17) a large diffusion length in
micrometers can be expected in 1/30 s even at low temperatures. It should be noted that this estimate is invalid for
materials that are maintained at high temperatures or under
irradiation, where high concentration of point defects, small
defect clusters, impurities, and even the fluctuation of atomic
arrangement are introduced. Such pre-existing defects or
lattice disorders work as the diffusion obstacles, enabling the
glissile-to-sessile transition. The diffusion distance depends
on the irradiation time and implantation fluence, which
directly relates to the lattice defect concentration. Continuous
irradiation allows subsequent ion impacts introducing the
inverse reaction, i.e., sessile-to-glissile transition, which is
affected by energy density of cascades, damage rate and
defect stability. According to the MD simulations,19) the
sessile-to-glissile transition requires approximately 27 meV
in the absence of a potential barrier.
Stability and Mobility of Interstitial-type Defect Clusters Generated from Displacement Cascades in Copper and Gold
Future prospects for the effects will be to disclose the
interactions between glissile defects and various kinds of
defects and their agglomerations. In particular, in low-doserate environments, such as nuclear reactor vessels and fusion
reactor components except for first wall, the glissile defects
diffuse in long range without showing any interactions with
the obstacles. These defects may accumulate at the grain
boundaries and precipitate surface resulting in deterioration
of the mechanical strength. These effects can be investigated
by MD simulations and other techniques, as well as by crosssectional TEM observations. An ideal simulation can be
performed with the materials such as copper, gold,30) and
coherent precipitates,31) along with an application of point
defect kinetics under electron irradiations32) and under
displacement cascade conditions.33)
4.
Conclusions
Defect clusters corresponding to displacement cascades
were observed by in-situ TEM observations in copper and
gold under irradiations of 100-keV Cþ , 240-keV Cuþ , 600keV Kr2þ , and 900-keV Xe3þ ions at temperatures ranging
from 573 to 823 K. In addition to the single contrast features
associated with the cascades, the multi-contrast features
arising from subcascades and glissile defects were observed.
The characteristic behaviors of glissile defects are their
anisotropic distributions, fluence-dependent distance, ionmass-dependence, and relatively short lifetime (less than
several thirtieths of a second) compared with the vacancy
loops and SFTs. In addition to the cascade-driven glissile
defects, sessile defects were also transformed into glissile
ones under irradiations with ions. This behavior is characterized by a one-dimensional back-and-forth motion and
spontaneous long range diffusion. These observations suggest that the sessile-to-glissile transition and the diffusion
along the crowdion directions as well as glissile-to-sessile
transition through interaction with certain diffusion obstacles
are invisible during microscopy. Inter-subcascade and intercascade events were observed at higher formation rates of
glissile defects for heavy ion irradiations and as non-linear
relationships between the formation rate and ion flux.
Acknowledgements
One of authors (HA) is grateful to Drs. S. J. Zinkle and R.
E. Stoller of the Oak Ridge National Laboratory, USA, for
their enthusiastic interest in the work and numerous discussions.
439
REFERENCES
1) T. Diaz de la Rubia and M. W. Guinan: J. Nucl. Mater. 174 (1990) 151–
157.
2) R. S. Averback, H. Hsieh, T. Diaz de la Rubia and R. Benedek: J. Nucl.
Mater. 179–181 (1991) 87–93.
3) A. Taylor, C. W. Allen and E. A. Ryan: Nucl. Instrum. Methods B24/
25 (1987) 598–602.
4) C. Kinoshita, H. Abe, K. Fukumoto, K. Nakai and K. Shinohara:
Ultramicroscopy 39 (1991) 205.
5) H. Abe, H. Naramoto, K. Hojou and S. Furuno: JAERI-Research 96047.
6) H. Sakaida, N. Sekimura and S. Ishino: J. Nucl. Mater. 179–181 (1991)
928–930.
7) K. Fukumoto, C. Kinoshita, H. Abe, K. Shinohara and M. Kutsuwada:
J. Nucl. Mater. 179–181 (1991) 935–938.
8) S. Ishino, N. Sekimura, H. Sakaida and Y. Kanzaki: Mater. Sci. Forum
97–99 (1992) 165–182.
9) I. M. Robertson, J. S. Vetrano, M. A. Kirk and M. L. Jenkins: Philos.
Mag. A63 (1991) 299–318.
10) S. Ishino, N. Sekimura, K. Hirooka and T. Muroga: J. Nucl. Mater.
141–143 (1986) 776–780.
11) S. Ishino, N. Sekimura and T. Muroga: Mater. Sci. Forum 15–18 (1987)
1105–1110.
12) H. Abe, C. Kinoshita and K. Nakai: J. Nucl. Mater. 179–181 (1991)
917–920.
13) L. M. Howe and M. H. Rainville: Nucl. Instrum. Methods B19/20
(1987) 61–66.
14) M. Kiritani: J. Nucl. Mater. 251 (1997) 237–251.
15) K. Arakawa, S. Arai, H. Orihara, K. Ono and M. Kiritani: J. Electron
Microscopy 51 (2002) S225–S229.
16) R. E. Stoller, G. R. Odette and B. D. Wirth: J. Nucl. Mater. 251 (1997)
49–60.
17) Yu. N. Osetsky, D. J. Bacon, A. Serra, B. N. Singh and S. I. Golubov:
J. Nucl. Mater. 276 (2000) 65–77.
18) Yu. N. Osetsky, D. J. Bacon, A. Serra, B. N. Singh and S. I. Golubov:
Philos. Mag. 83 (2003) 61–91.
19) H. Abe, N. Sekimura and Y. Yang: J. Nucl. Mater. 323 (2003) 220–228.
20) K. Niwase: private communication.
21) K. Niwase and H. Abe: Mater. Trans. 43 (2002) 646–649.
22) H. Abe, H. Naramoto, A. Iwase and C. Kinoshita: Nucl. Instrum.
Methods B127–128 (1997) 681–684.
23) J. F. Ziegler, J. P. Biersack and U. Littmark: The Stopping and Range of
Ions in Solids, (Pergamon Press, New York, 1985).
24) M. L. Jenkins and M. A. Kirk: Characterization of Radiation Damage
by Transmission Electron Microscopy, (Inst. of Physics Publ., 2001)
p. 11.
25) M. L. Jenkins and M. A. Kirk: Characterization of Radiation Damage
by Transmission Electron Microscopy, (Inst. of Physics Publ., 2001)
p. 115.
26) R. S. Walker and D. A. Thompson: Rad. Eff. 37 (1978) 113–120.
27) N. Sekimura: private communication.
28) L. E. Rehn and P. R. Okamoto: Nucl. Instrum. Methods, Phys. Res. B39
(1989) 104–113.
29) H. Abe: Dr. Eng. Thesis, Kyushu University (1993).
30) Y. Fujino, Y. Igarashi and S. Nagata: Phys. Rev. B63 (2001) 100101.
31) R. W. Knoll: Ph D. Thesis, Univ. Wisconsin-Madison (1981).
32) N. Yoshida and M. Kiritani: J. Phys. Soc. Jpn. 35 (1973) 1418–1429.
33) T. Muroga: Rad. Eff. Defects in Solids 144 (1998) 105–118.