Formation of CeSi2 on the Si surface upon high current pulsed Ce-ion implantation

Article abstract of DOI:10.1016/S0925-8388(03)00679-0

Cerium-ion implantation was conducted to synthesize Ce-disilicide films on silicon wafers, using a metal vapor vacuum arc ion source. The continuous CeSi₂ films were directly obtained at relatively low temperature with neither external heating nor post-annealing and the surface morphology varied with the variation of the implantation parameters. The formation mechanism of the CeSi₂ phase is also discussed in terms of the temperature rise caused by ion beam heating and the ion dose in the far-from-equilibrium process of high current pulsed Ce-ion implantation.

Full text of DOI:10.1016/S0925-8388(03)00679-0

Journal of Alloys and Compounds 365 (2004) 240–245
Formation of CeSi2 on the Si surface upon high current
pulsed Ce-ion implantation
X.Q. Cheng, X.J. Tang, B.X. Liu∗
Advanced Materials Laboratory, Department of Material Science and Engineering, Tsinghua University, Beijing 100084, China
Received 2 December 2002; received in revised form 2 June 2003; accepted 2 June 2003
Abstract
Cerium-ion implantation was conducted to synthesize Ce-disilicide films on silicon wafers, using a metal vapor vacuum arc ion source. The
continuous CeSi2 films were directly obtained at relatively low temperature with neither external heating nor post-annealing and the surface
morphology varied with the variation of the implantation parameters. The formation mechanism of the CeSi2 phase is also discussed in terms of
the temperature rise caused by ion beam heating and the ion dose in the far-from-equilibrium process of high current pulsed Ce-ion implantation.
© 2003 Elsevier B.V. All rights reserved.
Keywords: Thin films; Rare earth compounds; Microstructure; X-ray diffraction
1. Introduction
were typically dominated with heavy pits, which in turn
would have a detrimental effect on the electronic perfor-
mance [9]. In recent years, IBS technique has also been
employed to synthesize the RE metal silicides because of its
compatiblity with the Si planar technology. In general, the
IBS technique using the conventional implanters consists
of a high-dose ion implantation at a medium temperature
of 350–450 ◦C and a post-annealing conducted at a tem-
perature as high as 1000 ◦C. Subsequently, the RE metal
silicides could be successfully synthesized under the chan-
neling conditions [10–12]. It is possible to prevent the prob-
lem due to high reactivity of RE metals that may degrade
the quality of the surface, silicide/Si interface, and silicide
layers. Nonetheless, as the conventional implanters could
provide weak metal ion current of several pA/cm2 order,
high dose implantation of the order of 1017–1018 ions/cm2
was time consuming and lasted, at least, for several hours.
It is known that a new ion source, namely the metal va-
por vacuum arc (MEVVA) ion source, was invented in the
mid-1980s [13], and the source is capable of providing al-
most all the metal ion species with a very high current den-
sity up to 100 ␮A/cm2, which is at least 1–2 orders of
magnitude higher than that available in the conventional im-
planters. Comparatively, the higher dose ion implantation
can be completed within less time very effectively at high
current metal ion beam. Because of its intense ion current,
metal-ion implantation can cause a significant temperature
Metal silicides have been widely used in very large scale
integration (VLSI) technology to form the uniform and
thermally stable contacts, interconnects and low-resistivity
gates for high-speed and low-power-loss applications [1,2].
Recently rare-earth (RE) metal silicides have attracted con-
siderable attention, because they can form the lowest Schot-
tky barrier height (0.3–0.4 eV) on n-type silicon surface, and
because they are also excellent candidate materials of fab-
ricating the advanced infrared detectors [3–5]. Accordingly,
various techniques have been extensively developed to syn-
thesize RE metal silicides, such as solid-state reaction (SSR)
and ion beam synthesis (IBS) using conventional implanters.
Solid-state reaction was firstly employed to synthesize
the RE metal silicides. It was found that the interaction
between the RE metals and single crystalline silicon wafers
frequently behaves a ‘critical temperature’ phenomenon,
i.e., below the critical temperature (320–350 ◦C), the inter-
action was sluggish, whereas above the critical temperature,
the interaction was explosive [6–8]. As a result, the SSR
between the RE metal layer and Si could not be well con-
trolled and the RE metal silicide layers obtained by SSR
∗
Corresponding author. Fax: +86-10-6277-1160.
E-mail address: dmslbx@tsinghua.edu.cn (B.X. Liu).
0925-8388/$ – see front matter © 2003 Elsevier B.V. All rights reserved.
doi:10.1016/S0925-8388(03)00679-0

X.Q. Cheng et al. / Journal of Alloys and Compounds 365 (2004) 240–245
241
rise in the Si surface, resulting in a simultaneous thermal
annealing of the Si substrate by ion beam itself. It was
therefore possible to synthesize the metal silicides by us-
ing high current MEVVA ion source with neither external
heating during implantation nor post-annealing. A fast-step
technique of the MEVVA ion source has successfully been
employed by the authors’ group to synthesize the impor-
tant metal silicides as, such as C54-TiSi2, ␤-FeSi2, CoSi2,
NiSi2 and ZrSi2 [14–18]. Currently, the MEVVA ion source
works in a pulse mode, i.e., the current density in a pulse is
always the same and the increasing of the average current
densities is reached by increasing the pulse number within a
unit time. The physical process of the metal atoms dynami-
cally launched into the Si lattice in IBS using MEVVA ion
source should be obviously different from that of the ther-
mal diffusion and atom migration which emerged in SSR,
therefore, the temperature of forming the metal silicide by
high current MEVVA ion implantation can probably be con-
siderably lower than that required in SSR [19]. Note that
an effective implantation temperature here is defined as the
actual temperature rise of the Si wafer caused by the ion
beam heating, which could be directly measured, e.g., by a
thermal-couple attached on Si surface. We therefore investi-
gate, in the present study, the possibility of directly forming
the Ce-disilicide layers on Si wafers by high current Ce-ion
implantation and report, in this paper, the experimental ob-
servations concerning the formation of the CeSi2 layers on
Si wafers, the effects of ion current as well as dose on the
CeSi2 phase formation and present a brief discussion of the
formation mechanism of the CeSi2 upon MEVVA ion im-
plantation.
pulsed metal-ion beam and the time period of one pulse was
1.2 ms, which was much shorter than the interval (0.3–1.0 s)
between two consecutive pulses. Consequently, during high
current ion implantation, the temperature rise of the Si sub-
strate increased rapidly upon one pulse shot and then de-
creased until the next pulse. After a number of pulses, the
temperature rise of the Si substrate reached a balanced value,
which was actually an average or saturated value during im-
plantation. In high current ion implantation experiments, ad-
justing the ion current density could vary the temperature
rise of the Si substrate and the effect of temperature on the
Ce-silicide formation will be discussed later [20]. Since the
cathode of the MEVVA ion source was made of Ce with a
purity of 99.99 wt%, the purity of the extracted Ce ion beam
was considered to be about 99.99 wt%. In order to perform
various measurements, at least four samples were prepared
for each specific set of implantation parameters.
X-ray diffraction (XRD) was performed to identify the
crystalline structure of the formed Ce-silicides after MEVVA
ion implantation by a D/max-RB diffractometer operated
with a Cu radiation of wavelength 1.5418 Å at 40 kV and
120 mA. A step-scan method was adopted with 0.02◦ per
step and 3 s stopping time at each step. The X-ray beam spot
was around 0.5 × 1.0 cm2. Rutherford backscattering spec-
trometry (RBS) was employed to measure the depth profiles
of the implanted Ce-ions in the Si wafers with 2.0 MeV
He ions at a 165◦ scattering angle. The ion beam spot was
about 2 × 2 mm2. A scanning electron microscope (SEM)
was used to observe the surface and cross-sectional mor-
phology of the formed Ce-silicide layers on the Si wafers
with an acceleration voltage of 10 kV. To obtain reasonable
statistics, each measurement was repeated at least twice.
2. Experimental procedure
3. Results and discussion
The silicon wafers used in this study were n-type Si (100)
and Si (111) with a resistivity of 2–4 and 8–10 ꢀ·cm. The
wafers were cut into 1 × 1-cm2 samples. The samples were
cleaned by a standard chemical procedure and then dipped
in a dilute HF solution, followed by a rinse in deionized wa-
ter. The cleaned samples were then loaded onto a steel-made
sample holder in the target chamber of the MEVVA im-
planter operated at an extract voltage of 45 kV. The vac-
uum level of the MEVVA implanter was of 2 × 10−3 Pa.
During implantation, no deliberate heating was employed
for the samples. The samples were implanted with a tilted
angle of 7o to minimize channeling effects. As the im-
plantation system has no analysis magnet, the extracted
Cerium ions are multiple charged and have been analyzed to
consist of 3%Ce+, 83%Ce2+, 14%Ce3+, respectively. The
samples were implanted with the current densities varying
from 26.4 to 70.4 ␮A/cm2 to the nominal doses ranging
from 1 × 1017 to 4 × 1017 ions/cm2. A thermocouple was
placed on a non-implanted area of the Si surface to mea-
sure the effective temperature rise of the Si substrate dur-
ing Ce-ion implantation. The MEVVA implanter provided a
3.1. Formation of the CeSi2 phase
We now present the experimental results of CeSi2 forma-
tion upon high current Ce-ion implantation into Si surface
using a MEVVA ion source. Since the results obtained in the
present study for the Si (100) and Si (111) substrates were
similar, for the Si (100) samples, implanted with three differ-
ent current densities, i.e., 26.4, 52.8 and 70.4 ␮A/cm2, to a
fixed dose of 2 × 1017 ions/cm2, the corresponding XRD pat-
terns of the obtained Ce–Si phase are shown in Fig. 1. From
the patterns, one can clearly see that the tetragonal CeSi2
phase was formed in all of the above cases. The temperature
rises of the Si wafers under three different ion current densi-
ties of 26.4, 52.8 and 70.4 ␮A/cm2, were measured to be of
295, 380 and 430 ◦C, respectively, with a measurement er-
ror of 10 ◦C. These directly measured temperatures can be
considered as the effective formation ones of the Ce-silicide
phase under the respective conditions of ion implantation.
We now discuss the effect of ion current density on the for-
mation of the CeSi2 phase upon a MEVVA ion implantation.

242
X.Q. Cheng et al. / Journal of Alloys and Compounds 365 (2004) 240–245
the diffraction peaks shown in Fig. 1(c) are not stronger than
those in the Fig. 1(b), indicating that the crystalline struc-
ture was not further improved. Meanwhile, no strong texture
appears and it could be not easy to grow up at higher forma-
tion temperature. These results indicate that the CeSi2 phase
could be directly obtained at the formation temperature of
380 ◦C with well-crystallized structure on the Si wafers by
single-step ion implantation using the MEVVA ion source.
We also study the effect of ion dose on the CeSi2 phase
formation using the MEVVA ion source. For example, a
typical Si (111) sample was implanted by Ce-ion with a
fixed ion current density of 52.8 ␮A/cm2, corresponding
to a formation temperature of 380 ◦C, at ion doses ranging
from 1 × 1017 to 4 × 1017 ions/cm2. At the lowest dose of
1 × 1017 ions/cm2 in this experiment, the XRD pattern for
the Si (111) wafer was obtained as shown in Fig. 2(a). One
can see that most diffraction peaks reflected from CeSi2
phase appear in the spectrum. In Fig. 2(b), with increasing
the ion dose up to 2 × 1017 ions/cm2, the peaks are a little
sharper than those in Fig. 2(a). Further increasing the ion
dose up to 4 × 1017 ions/cm2, the XRD pattern in Fig. 2(c)
shows more diffraction peaks of the CeSi2 phase with a
strong (004) peak. It is therefore concluded that the opti-
mal experimental parameters for synthesizing the continu-
ous CeSi2 layers on Si (111) wafers by MEVVA ion im-
plantation are an ion current density of about 52.8 ␮A/cm2,
corresponding to a formation temperature of approximately
380 ◦C, and an ion dose around 2–4 × 1017 ions/cm2.
RBS measurement can identify the composition as well
as the stoichiometry of the formed CeSi2 layers. Fig. 3 is a
random RBS spectrum of the CeSi2 layer formed on Si (111)
after the implantation with Ce-ion current density of 52.4
␮A/cm2 to a dose of 4 × 1017 ions/cm2, corresponding to a
formation temperature of 380 ◦C. One can see that a little
plateau signal just behind the Si leading edge appears near
the Si surface and the Ce signal is relatively sharp, suggesting
that the formed layer is quite thin. As there is no apparent
tail at the low-energy edge of the Ce signal, the interface
between the formed CeSi2 layers and the Si substrate might
be relatively sharp, which will be further discussed together
with the morphology observed from cross-sectional SEM
examination.
Fig. 1. XRD patterns of Si (100) wafers implanted by Ce-ion with various
current densities to a fixed dose of 2 × 1017 ions/cm2. (a) 26.4 ␮A/cm2,
(b) 52.8 ␮A/cm2, and (c) 70.4 ␮A/cm2.
According to the nominal implantation dose of 4 × 1017
ions/cm2, the average range of the Ce-ion deduced from the
TRIM program was about 48 nm, which was considered as
the thickness of the Ce+Si mixture layer [21]. From the ex-
perimental spectrum, the thickness of the formed CeSi2 was
estimated to be about 40 nm, which was much thinner than
the former layer [22]. Such a difference could be attributed
to the fact that a large portion of the implanted Ce atoms
sputtered off from the surface upon interaction of high cur-
rent Ce-ion with Si substrate. It was estimated that the ac-
tual Ce dose retained in the sample is calculated to be about
8.2 × 1016 ions/cm2, which is up to the implanted saturation
of the Si substrate [23]. Furthermore, the composition pro-
file of the Ce atoms deduced from the spectrum confirmed
One sees from Fig. 1(a) that at a current density of 26.4
␮A/cm2, i.e., at a formation temperature of 295 ◦C, there
are only two weak diffraction peaks (004) and (112) from
the CeSi2 phase. With increasing the ion current density to
52.8 ␮A/cm2, i.e., at the formation temperature of 380 ◦C,
the crystalline structure of the CeSi2 phase was consider-
ably improved, as shown by the XRD pattern in Fig. 1(b).
All the diffraction peaks of CeSi2 phase obviously appear,
although they are very weak except a strong and sharp (004)
peak, indicating that an intense texture has emerged. This
result shows the possibility of preferential growth of CeSi2
on Si (100). Further increasing the ion current density up to
70.4 ␮A/cm2, i.e., at the formation temperature of 430 ◦C,

X.Q. Cheng et al. / Journal of Alloys and Compounds 365 (2004) 240–245
243
to a Gaussian distribution and an average stoichiometry of
the formed Ce:Si was determined to be about 1:2.0 with an
error of 2.0% at a distance of about 10 nm from the surface,
which corresponded quite closely to that of the equilibrium
CeSi2 phase. It is believed that the irradiation can enhance
diffusion of the Ce atoms to a great depth into the Si wafer,
resulting in that the Ce concentration at a distance could be
small than that in the CeSi2 phase.
3.2. Surface and cross-sectional morphology of
the CeSi2 layers
In this present study, SEM observations revealed some
interesting features of the CeSi2 layers obtained on the Si
(111) surface by high current pulsed Ce-ion implantation.
Fig. 3. RBS spectrum of the Si (111) sample with Ce-ion implantation:
52.8 ␮A/cm2, 4 × 1017 ions/cm2.
Fig. 4(a)–(c) shows a set of SEM morphologies observed
from some samples with various effective temperatures dur-
ing MEVVA ion implantation at the fixed dose of 2 × 1017
ions/cm2. In Fig. 4(a), at the lowest temperature of 295 ◦C,
a network pattern consisting of CeSi2 phase (the bright
regions) was observed on the Si surface, suggesting that
this morphology probably corresponds to an early growth
shape of CeSi2 grains. When the temperature was increased
to 380 ◦C, the CeSi2 grains grew up and embed in the Si
substrate, as shown in Fig. 4(b). Further increasing the
temperature up to 430 ◦C, the grain size of the CeSi2 layer
in Fig. 4(c) is even finer than that in Fig. 4(b), which
was because that the higher the temperature the faster the
crystals grow and the growing time would decrease when
the current density was increased. Consequently, it was a
high formation temperature yet with an adequate grow-
ing time that resulted in a dense aggregated morphology
shown in Fig. 4(c). To study the effect of ion dose, another
Si (111) sample was implanted with a current density of
52.8 ␮A/cm2 to a dose of 1 × 1017 ions/cm2 and the ob-
served morphology is shown in Fig. 4(d). Compared with
the case shown in Fig. 4(b), which was twice the dose in
Fig. 4(d), one can see that many CeSi2 grains did grow
larger or upright in a cylindrical shape when the ion dose
was increased. It should be noted that the implantation to
an additional dose could prolong the reaction time between
the Ce ions and the Si substrate, and more Ce atoms was
added into the Ce+Si mixture to increase the CeSi2 grain
size.
In order to examine the continuity of the formed CeSi2
layers, the cross-sectional SEM image was observed for the
Si (111) sample with a current density of 52.8 ␮A/cm2 to a
dose of 2 × 1017 ions/cm2. Fig. 5 exhibits the cross-sectional
SEM morphology, which clearly shows that the formed
CeSi2 layer is continuous, and the interface between the
CeSi2 layers and the Si substrate is relatively sharp.
Fig. 2. XRD patterns of Si (111) wafers implanted by Ce-ion with various
doses to a fixed current density of 52.8 ␮A/cm2. (a) 1 × 1017 ions/cm2,
(b) 2 × 1017 ions/cm2, and (c) 4 × 1017 ions/cm2.

244
X.Q. Cheng et al. / Journal of Alloys and Compounds 365 (2004) 240–245
Fig. 5. Cross-sectional SEM image of the formed CeSi2 layer implanted
with Ce-ions with a current density of 52.8 ␮A/cm2 to a dose of 2 × 1017
ions/cm2.
at the termination of the atomic collisions and according to
the atomic collision theory, it lasts only for extremely short
time (10−10 to 10−9 s), which allows only simple structured
alloy phase to nucleate and grow. As mentioned above, the
MEVVA ion source provided a pulsed metal-ion beam. In
the present study, the width of the pulse was 1.2 ms and the
time interval within two consecutive pulses was 0.3–1.0 s,
which was much longer than the relaxation time period after
atomic collision.
Accordingly, the formation mechanism of the CeSi2 phase
upon high current Ce-ion implantation was thought to pro-
ceed as follows. For simplicity, the process was divided into
three steps for each pulsed Ce-ion implantation. Firstly, the
high-energy Ce-ions (on the order of keV) were dynami-
cally implanted into the Si lattice to trigger drastic atomic
collisions among the Ce and Si atoms. At the termination
of Ce-ion implantation and atomic collisions, a highly en-
ergetic Ce–Si mixture was obtained and it was most likely
in a disordered state. In the second step of relaxation period
immediately following the atomic collision, the highly en-
ergetic Ce–Si mixture should relax towards equilibrium by
some atomic re-arrangement. In the Ce–Si system, the tetrag-
onal CeSi2 phase is the most stable equilibrium phase hav-
ing the lowest free energy. Consequently, some crystalline
CeSi2 grains should be formed during the relaxation period.
The third step is defined as the time period from the end
of the above relaxation to the next implantation pulse and
during the third step, the already crystallized CeSi2 grains
should undergo a random walk and organize themselves to
form clusters. If the CeSi2 clusters can grow into a suffi-
ciently large size before the arrival of the next implantation
pulse, they are not destroyed completely by the impact of
the incoming Ce ions and can remain in the Si surface layer.
When the next pulse arrives in, some new CeSi2 grains could
be formed and the growth of the previously retained CeSi2
clusters could continue to proceed. This growth process can
be named intermittent cluster diffusion limited aggregation.
The continuation of the above process results in formation
of a continuous layer, consisting of CeSi2 grains that even-
tually cover the entire Si surface, which was in agreement
with the result in Fig. 5.
Fig. 4. Surface morphologies of Si (111) wafers implanted by Ce-ions
with various conditions. (a) 26.4 ␮A/cm2, 2 × 1017 ions/cm2, (b) 52.8
␮A/cm2, 2 × 1017 ions/cm2, (c) 70.4 ␮A/cm2, 2 × 1017 ions/cm2, (d) 52.8
␮A/cm2, 1 × 1017 ions/cm2.
3.3. Formation mechanism of the CeSi2 layers
We now discuss the formation mechanism of the CeSi2
phase obtained by MEVVA ion implantation. It is commonly
known that the formation of an alloy phase upon energetic
ion implantation could be considered in a two-step process.
In the first step, atomic collisions are triggered by the irradi-
ating ions and a great number of atoms in the target material
together with the irradiating ions are in dynamic motion.
During this step, the structure of any desired alloy phase
cannot be fixed. The second step named relaxation begins

X.Q. Cheng et al. / Journal of Alloys and Compounds 365 (2004) 240–245
245
In Ce-ion implantation, the dynamic process that the
References
energetic Ce ions are implanted into the Si lattice did not
require the Si substrate to be at an elevated temperature.
However in SSR of deposited metal/Si for the formation of
metal silicides, it was necessary to heat the Si substrate to
a certain elevated temperature in order to thermally activate
the metal atoms to diffuse into the Si substrate. The high
current Ce-ion implantation simultaneously caused a signif-
icant temperature rise of the Si substrate, which helped in
forming the CeSi2 phase. These two effects made it possible
to form the CeSi2 phase at a considerably lower temper-
ature, which is a process mainly depending on thermal
activation and inter-diffusion. In addition, the well-known
radiation enhanced diffusion (RED) could also increase the
interaction between the Ce atoms and the Si substrates with
various temperatures, which could aid in promoting the
formation of the CeSi2 phase as well as in decreasing the
effective temperature upon MEVVA ion implantation.
[1] F.A. d’Avitaya, A. Perio, J.C. Oberlin, Y. Campidelli, J.A.
Chroboczek, Appl. Phys. Lett. 54 (1989) 2198.
[2] F.A. d’Avitaya, P.A. Badoz, Y. Campidelli, J.A. Chroboczek, J.Y.
Duboz, A. Perio, Thin Solid Films 184 (1990) 283.
[3] J.A. Knapp, S.T. Picraux, Appl. Phys. Lett. 48 (1986) 466.
[4] K.N. Tu, R.D. Thompson, B.Y. Tsaur, Appl. Phys. Lett. 38 (1981)
626.
[5] H. Norde, J. deSousa Pires, F. d’Heurle, F. Pesavento, S. Petersson,
P.A. Tove, Appl. Phys. Lett. 38 (1981) 865.
[6] J.E.E. Baglin, F.M. d’Heurle, C.S. Petersson, Appl. Phys. Lett. 36
(1980) 594.
[7] R.D. Thompson, B.Y. Tsaur, K.N. Tu, Appl. Phys. Lett. 38 (1981)
535.
[8] J.E.E. Baglin, F.M. d’Heurle, C.S. Petersson, J. Appl. Phys. 52 (1981)
2841.
[9] S.S. Lau, C.S. Pai, C.S. Wu, T.F. Kuech, B.X. Liu, Appl. Phys. Lett.
41 (1982) 77.
[10] A.E. White, K.T. Short, R.C. Dynes, J.P. Garno, J.M. Gibson, Appl.
Phys. Lett. 50 (1987) 95.
[11] M.F. Wu, A. Vantomme, H. Pattyn, G. Langouche, Appl. Phys. Lett.
67 (1995) 3886.
[12] M.F. Wu, A. Vantomme, S. Hogg, H. Pattyn, G. Langouche, Appl.
Phys. Lett. 72 (1998) 2412.
4. Conclusion
[13] I.G. Brown, J.E. Gavin, R.A. MacGill, Appl. Phys. Lett. 47 (1985)
358.
[14] D.H. Zhu, K. Tao, F. Pan, B.X. Liu, Appl. Phys. Lett. 62 (1993)
2356.
[15] B.X. Liu, D.H. Zhu, H.B. Lu, F. Pan, K. Tao, J. Appl. Phys. 75
(1994) 3847.
[16] D.H. Zhu, Y.G. Chen, B.X. Liu, Nucl. Instrum. Methods B101 (1995)
394.
We have shown that the equilibrium CeSi2 phase could
directly be synthesized on Si surfaces by a single-step pro-
cess of high current Ce-ion implantation using MEVVA
ion source and that the formed CeSi2 layers with a
well-crystalline structure as well as a smooth surface mor-
phology could be obtained under an optimized condition in
high current Ce-ion implantation.
[17] K.Y. Gao, B.X. Liu, Nucl. Instrum. Methods B132 (1997) 68.
[18] K.Y. Gao, H.N. Zhu, B.X. Liu, Nucl. Instrum. Methods B140 (1998)
129.
[19] H.N. Zhu, K.Y. Gao, B.X. Liu, Phys. Rev. B62 (2000) 1647.
[20] D.H. Zhu, B.X. Liu, J. Appl. Phys. 77 (1995) 6257.
[21] J.F. Ziegler, J.P. Biersack, TRIM (1992) Pergamon PressNew York.
[22] M. Mayer, SIMNRA, V4.0, Max Planck Institute 1997.
[23] G.W. Reynolds, A.R. Knudson, C.R. Gossett, Nucl. Instrum. Methods
182/183 (1981) 179.
Acknowledgements
The financial aids for this study from the National Nat-
ural Science Foundation of China, the Ministry of Science
and Technology of China (Grant No. G2000067207-1) and
the Administration of Tsinghua University are gratefully ac-
knowledged.