Cobalt metallorganic chemical vapor deposition and formation of epitaxial CoSi2 layer on Si(100) substrate

Article abstract of DOI:10.1149/1.1391999

An epitaxial CoSi₂ layer was grown on Si(100) substrate by the diffusion of Co from a cobalt-carbon film without the use of an interlayer, which is usually required between the Si and cobalt layers for the formation of epitaxial CoSi₂ layers. Co-C and pure Co layers were deposited by metallorganic chemical vapor deposition using the Co precursor cyclopentadienyl dicarbonyl cobalt, Co(η⁵-C₅H₅(CO)₂, and cobalt carbonyl, Co₂(CO)₈, at 350 and 200 °C respectively. The CoSi₂ layer was epitaxially grown on Si(100) substrate from Co-C by ex situ rapid thermal annealing at 800 °C in N₂ ambient. However, the polycrystalline CoSi₂ layer was formed from pure Co film under the same annealing conditions. The supply of Co to the interface by diffusion in the Co-C film seems to be lower enough than in the pure Co film, resulting in an epitaxial CoSi₂ layer on Si(100) substrate.

Full text of DOI:10.1149/1.1391999

2
720
Journal of The Electrochemical Society, 146 (7) 2720-2724 (1999)
S0013-4651(98)10-013-7 CCC: $7.00 © The Electrochemical Society, Inc.
Cobalt Metallorganic Chemical Vapor Deposition and Formation of
Epitaxial CoSi Layer on Si(100) Substrate
2
Hwa Sung Rhee and Byung Tae Ahna,*,z
Department of Materials Science and Engineering, Korea Advanced Institute of Science and Technology, 373-1 Koosung-dong,
Yusung-gu, Taejon 305-701, Korea
An epitaxial CoSi layer was grown on Si(100) substrate by the diffusion of Co from a cobalt-carbon film without the use of an
2
interlayer, which is usually required between the Si and cobalt layers for the formation of epitaxial CoSi layers. Co–C and pure
2
Co layers were deposited by metallorganic chemical vapor deposition using the Co precursor cyclopentadienyl dicarbonyl cobalt,
5
Co(␩ -C H )(CO) , and cobalt carbonyl, Co (CO) , at 350 and 200ЊC, respectively. The CoSi layer was epitaxially grown on
5
5
2
2
8
2
Si(100) substrate from Co–C by ex situ rapid thermal annealing at 800ЊC in N ambient. However, the polycrystalline CoSi layer
2
2
was formed from pure Co film under the same annealing conditions. The supply of Co to the interface by diffusion in the Co–C
film seems to be lower enough than in the pure Co film, resulting in an epitaxial CoSi layer on Si(100) substrate.
2
©
1999 The Electrochemical Society. S0013-4651(98)10-013-7. All rights reserved.
Manuscript submitted October 5, 1998; revised manuscript received March 8, 1999.
5
In ultralarge-scale integrated circuit (ULSI) technology, metal sili-
cide interconnections in the source/drain and gate regions have been
used to reduce contact resistance and voltage drop. Among the possi-
compared to Co(␩ -C H )(CO) in low-pressure chemical vapor
5
5
2
13
deposition (LPCVD). In this work, we investigated the reaction
with Si(100) substrate and pure cobalt and cobalt-carbon (Co-C)
films deposited using MOCVD by one-step ex situ annealing.
ble silicide candidates, CoSi is considered an attractive contact mate-
2
rial for submicron devices because of its low resistivity and ease of
1
formation on narrow Si lines. In particular, an epitaxial CoSi layer
Experimental
2
is required for good thermal stability and compatibility with shallow
junction formation using silicide-as-doping-source (SADS). Howev-
Substrates used were p-type Si(100) substrates with 5-8 ⍀ cm
2
and Si covered with 300 nm SiO . p-Type Si(100) substrates were
2
er, growth of epitaxial CoSi on Si(100) has not been successful by
2
cleaned in H SO /H O solution, rinsed in deionized water, dipped
2
4
2 2
physical vapor deposition (PVD) of a Co film on Si(100) and a sub-
sequent annealing process due to the various epitaxial orientations of
in HF(1%), rinsed in deionized water, and immediately loaded into
5
an MOCVD reactor. The Co precursors Co (CO)8 and Co(␩ -
2
3,4
CoSi , such as CoSi (110), (100), and (221) on Si(100).
2
2
C H )(CO) (Strem Chemicals) were used without further purifica-
5
5
2
The initial reaction between Co and Si causes the formation of
tion. For the rapid thermal annealing (RTA), pure Co films of 40 nm
thickness were deposited from Co (CO) at 50 mTorr with no carri-
the polycrystalline Co Si and CoSi phases at temperatures ranging
2
2
8
from 450 to 550ЊC, followed by the formation of a polycrystalline
er gas. The temperature of the bubbler and substrate were 35 and
CoSi phase above 600ЊC. Skipping the formation of Co-rich phas-
2
200ЊC, respectively. Co-C films of 40 nm thickness were deposited
5
es such as Co Si and CoSi during the reaction between Co and Si is
2
from Co(␩ -C H )(CO) at 440 mTorr with 50 sccm H carrier gas
5
5
2
2
5
related to the epitaxial growth of CoSi . Therefore, the suppression
2
when the temperature of the bubbler and the substrate were 35 and
50ЊC, respectively. To prevent the reaction between Si substrate and
deposited Co in the reactor during the deposition, the temperature of
the substrate was suppressed below 400ЊC in each case. A Ti capping
layer of 20 nm thickness was deposited on the films by dc magnetron
sputtering to avoid the oxidation of Co during annealing. A capping
of the Co supply at increasing temperatures can lead to the forma-
3
tion of an epitaxial CoSi on Si(100) substrates. Recently, the titani-
2
6
um interlayer–mediated epitaxy (TIME) and oxide-mediated epi-
taxy (OME) methods have been reported for the epitaxial growth of
7
CoSi2 on Si(100). Besides interlayer-mediated epitaxy methods,
8
9,10
Co-Ti and Co-W alloys
have been used for the formation of epi-
layer such as TiN or Ti generally improves the uniformity of CoSi
2
1
8
taxial CoSi . It has been proposed that an epitaxial CoSi layer can
2
2
layers on Si substrate. Subsequent ex situ RTA was carried out at
00ЊC for 5 min in N ambient. The thicknesses of as-deposited Co
be formed by controlling the concentration of Co into Si(100) sub-
8
2
1
1
12
strates using a diffusion barrier such as Ti, Ge, Zr and chemical
oxide in the interlayer-mediated techniques. The supply of Co to the
films were determined with a stylus profilometer. The crystal struc-
ture and microstructure of the films were investigated using X-ray
diffraction (XRD) and transmission electron microscopy (TEM),
respectively. The composition of the films was analyzed by Auger
electron spectroscopy (AES).
interface by diffusion in amorphous Co-W alloy also can be sup-
9
pressed, resulting in an epitaxial CoSi on Si(100).
2
We are introducing a new method for the formation of an epitax-
ial CoSi layer on Si(100) without employing an interlayer between
2
Co and Si using Co films deposited by metallorganic chemical vapor
Results and Discussion
deposition (MOCVD). Co MOCVD using the precursor cyclopenta-
The cobalt films were deposited from Co (CO) on Si and Si
2
8
5
dienyl dicarbonyl cobalt, Co(␩ -C H )(CO) , and cobalt carbonyl,
5
5
2
covered with 300 nm SiO at temperatures between 150 and 300ЊC.
2
13-16
Co (CO) , has been reported but has not been established yet.
2
8
The deposition pressure was 50 mTorr without carrier gas and the
selectivity was not obtained as to substrates. Figure 1 shows the
deposition rate using the precursor Co (CO) on SiO substrate as a
Chemical vapor deposition (CVD) using a metallorganic precursor
often does not produce pure metal films at low temperatures due to
incomplete decomposition of the metal-carbon bond. As the metal-
carbon bond order increases in organometallic compounds, the car-
bon incorporation in deposited film increases due to the increased
2
8
2
function of substrate temperature. At 200ЊC the deposition rate has a
maximum and decreases with increasing temperatures above 200ЊC.
Similar results have already been reported for Co deposition from
17
1
5,16
strength of the metal-carbon bond. The metal-carbon bond strength
in cyclopentadienyl dicarbonyl is stronger than that in cobalt car-
Co (CO) in atmospheric pressure CVD.
Dormans et al. pro-
2
8
posed that the growth rate decreases above 200ЊC due to increasing
bonyl. Therefore, MOCVD using Co (CO) produces purer films
2
8
homogeneous gas-phase reaction of the precursor molecule with
1
4
increasing temperature.
Figure 2 shows the deposition rate using the precursor Co(␩5-
C H )(CO) on SiO substrate as a function of substrate tempera-
*
Electrochemical Society Active Member.
E-mail: btahn@cais.kaist.ac.kr
z
5
5
2
2
Downloaded on 2015-02-27 to IP 138.251.14.35 address. Redistribution subject to ECS terms of use (see ecsdl.org/site/terms_use) unless CC License in place (see abstract).

Journal of The Electrochemical Society, 146 (7) 2720-2724 (1999)
2721
S0013-4651(98)10-013-7 CCC: $7.00 © The Electrochemical Society, Inc.
Figure 3. XRD pattern of the as-deposited Co films from the precursor (a)
Figure 1. Deposition rate of Co films from the precursor Co (CO) as a func-
tion of temperature without carrier gas at 50 mTorr.
5
2
8
Co (CO) at 200ЊC and (b) Co(␩ -C H )(CO) at 350ЊC.
2
8
5
5
2
Figure 4 shows the AES depth profile of the as-deposited films
5
ture. Depositions with Co(␩ -C H )(CO) were carried out at tem-
5
5
5
2
from precursor (a) Co (CO) and (b) Co(␩ -C H )(CO) . The film
2
8
5
5
2
peratures between 350 and 600ЊC with 50 sccm H carrier gas. The
deposition pressure was 440 mTorr. Co(␩ -C H )(CO) has no
selectivity on Si and SiO substrates. The deposition rate in LPCVD
shows a thermally activated temperature dependence that is com-
monly observed in other CVD systems. The reaction-rate-limited
regime at low temperatures and the mass-transport-limited regime at
high temperatures are shown and their activation energy is 0.513 and
2
deposited from Co (CO) has a relatively pure Co depth profile with
2 8
5
5
5
2
a small concentration of carbon and oxygen. Carbon in the as-de-
posited film from Co (CO) is likely to be contaminated due to
2
2
8
atmospheric exposure. Co (CO) is a solid with a vapor pressure of
2
8
about 60 mTorr at room temperature. No additional reducing agent
should be necessary for decomposition of the carbonyl to pure metal.
Therefore, the deposition of Co from Co (CO) is only by pyrolysis
2
8
0
.156 eV, respectively.
14
without extra reductant
Figure 3 shows the XRD patterns of the as-deposited Co films
from precursor (a) Co (CO) at 200ЊC and (b) Co(␩ -C H )(CO) at
5
Co (CO) (g) r 2Co(s) ϩ 8CO(g)
[1]
2
8
2
8
5
5
2
3
50ЊC. The Co film deposited from Co (CO) was the ␣-cobalt with
2 8
The AES depth profile in Fig. 4b indicates that the as-deposited film
consists of about 50 atom % cobalt and about 50 atom % carbon. The
carbon incorporation (about 50 atom %) is attributed to the decom-
a strong peak of 44.5Њ (002). However, the XRD pattern of Co film
deposited from Co(␩ -C H )(CO) shows a broad peak at the center
of 44-45Њ, but no sharp peak was detected. This indicates that the as-
deposited Co film from Co (CO) was in the polycrystalline phase,
but the as-deposited Co film from Co(␩ -C H )(CO) was in the
X-ray amorphous phase.
5
5
5
2
5
position of Co(␩ -C H )(CO) due to the unstable cyclopentadienyl
5
5
2
2
8
radical and seemed to cause the formation of the X-ray amorphous
5
5
5
2
5
phase in Fig. 3b. The cyclopentadienyl group in Co(␩ -C H )(CO)
5
5
2
reacts with sufficient H to form stable and volatile hydrocarbon
2
1
4
which is removed through the gas phase
5
Figure 2. Deposition rate of Co films from the precursor Co(␩ -C H )(CO)
2
Figure 4. AES depth profiles of the as-deposited films from precursor (a)
5
5
5
as a function of temperature with H carrier gas at 440 mTorr.
Co (CO) and (b) Co(␩ -C H )(CO) .
2
2
8
5
5
2
Downloaded on 2015-02-27 to IP 138.251.14.35 address. Redistribution subject to ECS terms of use (see ecsdl.org/site/terms_use) unless CC License in place (see abstract).

2
722
Journal of The Electrochemical Society, 146 (7) 2720-2724 (1999)
S0013-4651(98)10-013-7 CCC: $7.00 © The Electrochemical Society, Inc.
1
Co(C H )(CO) (g) ϩ nH (g) → Co(s) ϩ 2CO(g) ϩ C H5ϩn(g)
5
5
2
2
5
2
[
2]
where n ϭ 0, 1, 3, 5. Co(␩ -C H )(CO) seems to be no longer
5
5
5
2
decomposed by reduction due to the low H pressure in LPCVD but
2
by thermal disproportion, leading to the unstable and nonvolatile
1
9
cyclopentadienyl radical such as dihydrofulvalence.
Figure 5 shows the XRD patterns of annealed samples for
Ti/Co/Si(100) from the precursor (a) Co (CO) (sample A) and (b)
2
8
5
Co(␩ -C H )(CO) (sample B) after RTA at 800ЊC for 5 min in N
5
5
2
2
ambient. A CoSi layer was formed on Si(100) for each sample after
2
annealing. Figure 5a shows the XRD peaks of CoSi (111), (200),
2
(220), and substrate Si(200). For the sample used with the pure Co
layer, the polycrystalline CoSi layer was formed on Si(100) sub-
2
strate. However, only CoSi (200) and Si(200) peaks are shown in
2
Fig. 2b. No other peaks of CoSi such as (111) and (220) can be
2
seen, indicating that a CoSi layer is formed on a Si(100) substrate
2
with a strong (100) preferred orientation from the Co-C layer.
The AES depth profiles of sample A and sample B are shown in
Fig. 6a and b, respectively. Oxygen was capped by a Ti capping layer
and was difficult to diffuse into the Co layer in each case. Figure 6a
shows that Co further diffused into the Si substrate from the pure Co
layer than from the Co-C layer in the same annealing time. On the
other hand, Fig. 6b shows three distinct layers on the Si substrate.
The top layer is a TiO layer where the oxygen comes from the envi-
ronment and reacted with Ti. The second layer is a carbon layer con-
sisting of C, Co, and Si. It can be seen that the annealing lowers the
content of Co atoms, while Si atoms are introduced into the Co-C
layer. The ratio of Si/Co in this layer is about 2, apparently satisfy-
Figure 6. AES depth profiles of the annealed samples for Ti/Co/Si(100) from
5
precursor (a) Co (CO) and (b) Co(␩ -C H )(CO) at 800ЊC for 5 min in N
2
8
5
5
2
2
ambient.
face of CoSi2/Si(100) shows that the diffracted spots from CoSi2
and Si almost coincide with each other, which means that the
CoSi2 layer has coherent (100) planes with the Si(100) substrate
and an epitaxial CoSi2 layer was grown on the Si(100) substrate.
Since both Si and CoSi2 are in [011] orientation, the diffracted pat-
terns between Si and CoSi2 are indistinguishable.4 Figure 7c
shows a broad-ring pattern from the TiO and C-Co-Si layers, indi-
cating that an amorphous layer was formed after annealing due to
the large amount of carbon incorporation. The C-Co-Si layer is a
multicomponent layer consisting of C and a small amount of Co
and Si, as shown in Fig. 6b.
,8
ing the stoichiometry of CoSi . The first and second layers are X-ray
2
amorphous phase, because no XRD peaks related with the layers are
detected in Fig. 5b. The third layer is the CoSi layer where the dif-
2
fusion of Co seems to stop.
Figure 7 shows (a) the cross-sectional TEM (XTEM) micro-
graph of an annealed sample using Co-C film at 800ЊC for 5 min
in N and (b) the selected-area diffracted patterns (SADP) at the
Figure 8 shows the XTEM micrographs (a) of an annealed sam-
ple using Co-C film at 800ЊC for 1 min in N2 and (b) of an
annealed sample using pure Co film at 800ЊC for 5 min in N2. In
the initial stage of reaction between Co-C and Si, the discrete
CoSi2 layers with {111} and (100) facets are locally formed in the
Si substrate, marked with arrows in Fig. 8a. The epitaxial CoSi2
layer seems to grow discontinuously on a Si substrate with both
2
interface of CoSi /Si and (c) at the TiO/C-Co-Si interface along the
2
<
011> zone axis. The dark TiO layer, gray C-Co-Si layer, and dark
CoSi layers are seen on Si substrate in Fig. 7a. The CoSi layer
2
2
has a unifrom thickness of about 50 nm with small {111} facets at
the CoSi /Si interface. In Fig. 7b, the diffracted pattern at the inter-
2
{
111} and (100) facets, because the {111} interface energy of
CoSi2 is lower than (100). For the plasma-enhanced CVD
PECVD) of TiCl , Saito et al. reported that Ti atoms diffuse into
(
4
Figure 7. XTEM image (a) of annealed samples for Ti/Co/Si(100) from pre-
5
Figure 5. XRD patterns of the annealed samples for Ti/Co/Si(100) from pre-
cursor Co(␩ -C H )(CO) at 800ЊC for 5 min in N ambient and (b) SADP
5
5
2
2
5
cursor (a) Co (CO)8 and (b) Co(␩ -C H )(CO) at 800ЊC for 5 min in N2
at the CoSi /Si interface and (c) around the interface between the carbon
2
5
5
2
2
ambient.
layer and TiO layer.
Downloaded on 2015-02-27 to IP 138.251.14.35 address. Redistribution subject to ECS terms of use (see ecsdl.org/site/terms_use) unless CC License in place (see abstract).

Journal of The Electrochemical Society, 146 (7) 2720-2724 (1999)
2723
S0013-4651(98)10-013-7 CCC: $7.00 © The Electrochemical Society, Inc.
5
Figure 8. XTEM image of annealed samples for Ti/Co/Si(100) from (a) Co(␩ -C H )(CO) at 800ЊC for 1 min in N ambient and (b) Co (CO) at 800ЊC for 5
5
5
2
2
2
8
min in N ambient.
2
strate.21,22 Also note that the concentration of diffused Co at the C-
the Si from the surface in the form of low-spatial-density Co atoms
in the initial growth stage. On the other hand, Reader et al. pro-
2
0
Co-Si/CoSi interface is held to a lower level of the concentration
2
posed that a thin epitaxial CoSi layer is formed on the Si substrate
2
than in CoSi , as seen in Fig. 6b. This means that the formation of
2
in the initial stage of reaction between Co and Si and then the new
Co-rich phases such Co Si and CoSi can be avoided during the
2
CoSi layer is epitaxially grown by the diffusion of Co from the
2
reaction. The supply of Co by diffusion in the Co-C film seemed to
9
amorphous Co-W alloy into CoSi . The precise behavior of the
2
be low enough to form an epitaxial CoSi layer.
2
epitaxial growth from the Co-C film in the initial stage of reaction
is not yet fully understood and deserves further investigation. Fig-
ure 8b shows the CoSi2 layer from pure Co film is formed on
Conclusions
Si(100) substrate with a rough interface. The CoSi layer is in a
An epitaxial CoSi2 layer was grown on Si(100) substrate by the
diffusion of Co from a Co-C film as deposited using MOCVD of
2
polycrystalline phase from the previous XRD result of Fig. 5a.
5
This nonuniform penetration of polycrystalline CoSi layer is not
Co(␩ -C5H5)(CO)2. The employment of an intermediate buffer
2
useful for application to ULSI.
A high-resolution TEM micrograph along the <110> zone axis
was taken to investigate the coherency between CoSi and Si layers.
layer, usually required between Si and pure Co, was eliminated in
this experiment. However, a polycrystalline CoSi2 layer was formed
on Si(100) substrate from pure Co film as deposited from Co2(CO)8
after annealing. In this work, the technique using Co-C film by
MOCVD is proposed as another method to form the epitaxial CoSi2
layer on Si(100) substrates.
2
The lattice structure image obtained is shown in Fig. 9. Complete
registry of the lattice can be seen across the interface, indicating that
CoSi is fully coherent with the Si substrate. The epitaxial CoSi
2
2
layer has coherent (100) planes and small (111) facet planes with the
Si(100) substrate from the results of Fig. 7 and 9. Therefore, a CoSi2
layer can be epitaxially grown on a Si(100) substrate using a Co-C
The Korea Advanced Institute of Science and Technology assisted in
meeting the publication costs of this article.
layer without an interlayer such as Ti.
The role of the Co-C layer in the formation of an epitaxial CoSi2
layer can be ascertained from the following points. The key to form
an epitaxial CoSi on Si(100) seems to be the suppression of the Co
2
supply into the Si substrate at increasing temperatures. If the supply
of Co to the silicide/Si interface is low enough and the temperature
is high enough, an epitaxial CoSi layer is likely to form because
2
the Co consumption rate is larger than the Co supply rate. The
chemical oxide in OME, or the Ti layer in TIME, serves as a diffu-
sion barrier with the Co diffusing through it, thereby decreasing the
Co concentration to a level where CoSi2 thermodynamically is
favored. The epitaxial growth of CoSi2 using amorphous Co-W
alloy is attributed to the limited supply of Co to the interface by dif-
fusion. In this experiment, the supply of Co into the Si substrate is
restrained in the Co-C film as-deposited using Co metallorganic
precursor so that the diffused Co is completely consumed at the
interface, as in the amorphous alloy case. Figure 6 shows that Co
from the Co-C film is relatively more difficult to diffuse into the Si
substrate than from the pure Co film in the same annealing time,
indicating that the supply of Co in Co-C film is more suppressed
than in pure Co film. It has been reported that carbon at the inter-
face of silicide/Si could block the Co diffusion path into the Si sub-
Figure 9. HRTEM image along the <011> zone axis of the annealed sample
at 800ЊC for 5 min in N2 ambient from precursor Co(␩ -C5H5)(CO)2.
5
Downloaded on 2015-02-27 to IP 138.251.14.35 address. Redistribution subject to ECS terms of use (see ecsdl.org/site/terms_use) unless CC License in place (see abstract).

2
724
Journal of The Electrochemical Society, 146 (7) 2720-2724 (1999)
S0013-4651(98)10-013-7 CCC: $7.00 © The Electrochemical Society, Inc.
References
K. Maex, Mater. Sci. Eng., R, 11, 53 (1993).
12. S. L. Zhang and J. Cardenas, Appl. Phys. Lett., 66, 258 (1995).
1
3. G. J. M. Dormans, J. Cryst. Growth, 108, 806 (1991).
1
.
.
.
.
1
4. G. J. M. Dormans, G. J. B. M. Meekes, and E. G. J. Staring, J. Cryst. Growth, 114,
2
3
4
E. C. Jones, N. W. Cheung, and D. B. Fraser, J. Electron. Mater., 24, 863 (1995).
S. M. Yalisove, R. T. Tung, and D. Loretto, J. Vac. Sci. Technol., A7, 1472 (1989).
J. R. Jimenez, L. J. Schowalter, L. M. Hsiung, K. Rajan, S. Hashimoto, R. D.
Thomson, and S. S. Lyer, J. Vac. Sci. Technol., A8, 3014 (1990).
R. Pretorius and J. W. Mayer, J. Appl. Phys., 81, 2448 (1997).
M. L. A. Dass, D. B. Fraser, and C. S. Wei, Appl. Phys. Lett., 58, 1308 (1991).
R. T. Tung, Appl. Phys. Lett., 68, 3461 (1996).
3
64 (1991).
1
5. M. E. Gross, K. Schnoeskranz, D. Brasen, and H. Luftman, J. Vac. Sci. Technol.,
B6, 1548 (1988).
16. G. A. West and K. W. Beeson, Appl. Phys. Lett., 53, 740 (1988).
17. R. Kaplan and N. Bottka, Appl. Phys. Lett., 41, 972 (1982).
18. R. T. Tung and F. Schrey, Appl. Phys. Lett., 67, 2164 (1995).
5
6
7
8
9
.
.
.
.
.
S. L. Hsia, T. Y. Tan, P. Smith, and G. E. McGuire, J. Appl. Phys., 70, 7579 (1991).
A. H. Reader, J. P. W. B. Duchateau, J. E. Crombeen, and E. P. Naburgh, Appl. Surf.
Sci., 53, 92 (1991).
19. E. F. Knights and H. C. Brown, J. Am. Chem. Soc., 11, 5284 (1968).
20. K. Saito and Y. Arita, J. Electrochem. Soc., 143, 3778 (1996).
21. R. A. Donaton, K. Maex, A. Vantomme, G. Langouche, Y. Morciaux, A. St. Amour,
and J. C. Sturm, Appl. Phys. Lett., 70, 1266 (1997).
1
1
0. S. Q. Wang and J. W. Mayer, J. Appl. Phys., 65, 1957 (1989).
1. K. Prabhakaran, K. Sumitomo, and T. Ogino, Appl. Surf. Sci., 117/118, 280 (1997).
22. H. S. Rhee, T. W. Jang, and B. T. Ahn, Appl. Phys. Lett., 74, 1003 (1999).
Downloaded on 2015-02-27 to IP 138.251.14.35 address. Redistribution subject to ECS terms of use (see ecsdl.org/site/terms_use) unless CC License in place (see abstract).