Silicide formation in Co/Si system investigated by depth-resolved positron annihilation and X-ray diffraction

Article abstract of DOI:10.1016/j.susc.2006.04.043

The transformation of Co/Si to CoSi₂/Si in the temperature range of 300-1170 K has been investigated using depth-resolved positron annihilation and Glancing incidence X-ray diffraction (GIXRD). The different silicide phases formed are identifie

Full text of DOI:10.1016/j.susc.2006.04.043

Surface Science 600 (2006) 2762–2765
www.elsevier.com/locate/susc
Silicide formation in Co/Si system investigated by depth-resolved
positron annihilation and X-ray diffraction
*
S. Abhaya, G. Venugopal Rao, S. Kalavathi, V.S. Sastry, G. Amarendra
Materials Science Division, Indira Gandhi Centre for Atomic Research, Kalpakkam 603 102, India
Received 28 December 2005; accepted for publication 28 April 2006
Available online 22 May 2006
Abstract
The transformation of Co/Si to CoSi2/Si in the temperature range of 300–1170 K has been investigated using depth-resolved positron
annihilation and Glancing incidence X-ray diffraction (GIXRD). The different silicide phases formed are identified from the experimental
positron annihilation characteristics, which are consistent with the GIXRD results. The present study clearly indicates the absence of
vacancy defects in the silicide overlayer.
ꢀ 2006 Elsevier B.V. All rights reserved.
Keywords: Cobalt silicides; Positron spectroscopy; X-ray diffraction
1. Introduction
paper, we report the results of depth-resolved positron anni-
hilation spectroscopy (PAS), which has been used to inves-
tigate the formation of silicide phases in Co/Si system as
well as to probe the presence or absence of vacancy defects
consequent to silicide phase formation. Glancing incidence
X-ray diffraction (GIXRD) has been used to obtain corrob-
orative evidence for the silicide phases.
Study of metal silicides has been a topic of intense basic
and technological interest because of their potential applica-
tions in semiconductor industry as Schottky barriers, gates,
Ohmic contacts, etc. [1,2]. The effect of silicide formation on
the generation and the removal of defects is an important is-
sue for better understanding of the mechanism concerning
silicide formation. Wen et al. [3], through electrical mea-
surements, reported that the implantation induced intersti-
tials in Si were totally eliminated through the formation of
TiSi2 by the injection of vacancies. Similarly, silicidation
reactions of metals on Si implanted with various species
were also investigated [4] and it was found that the complete
annihilation of implantation induced defects depends criti-
cally on the distance between the silicide/Si interface, the
location of the original amorphous/crystalline interface,
the annealing temperature, time etc. Thus, it is important
to study the defects and their influence on silicidation using
a defect characterization tool together with a complemen-
tary structural or phase characterization technique. In this
2. Experimental
Pure Co (99.99%) was vapor-deposited onto Si (100) wa-
fer at room temperature using a resistive evaporation tech-
nique in a vacuum of 1 · 10À6 Torr. Prior to the deposition,
the substrate was cleaned using the standard HF acid etch-
ing procedure, followed by an in-situ Ar-ion sputtering to
remove the native oxide layer. The thickness of the as-
deposited Co film was estimated to be ꢀ90 nm from the
DEKTAK measurement and 95 10 nm from the Ruther-
ford backscattering spectrometry (RBS). The wafer thus
obtained was cut into smaller pieces and annealed in the
temperature range 370–1170 K in steps of 100 K for a dura-
tion of 1 h each in a vacuum of 1 · 10À6 Torr. Depth-
resolved positron annihilation measurements were carried
out at room temperature on the annealed Co/Si samples
*
Corresponding author. Tel./fax: +91 44 27480081.
E-mail address: amar@igcar.gov.in (G. Amarendra).
0039-6028/$ - see front matter ꢀ 2006 Elsevier B.V. All rights reserved.
doi:10.1016/j.susc.2006.04.043

S. Abhaya et al. / Surface Science 600 (2006) 2762–2765
2763
using a magnetically guided positron beam system [5]. By
varying the positron beam energy (Ep) from 200 eV to
20 keV, positrons can be implanted at selective depth re-
gions of the sample. The positrons so implanted get thermal-
ized and annihilate with the electrons of the sample giving
out 511 keV c-rays. The Doppler broadened 511 keV c-
ray spectrum is recorded using a HPGe detector, having
an energy resolution of 1.4 keV at 662 keV. From the mea-
sured Doppler broadened spectrum, a defect sensitive line-
shape S-parameter and a complimentary W-parameter are
deduced at each Ep. The S-parameter is defined as the
ratio of counts in the central region (511 1 keV) to the
total counts under the photo peak (511 10 keV) and
the W-parameter is defined as the ratio of counts in the wing
regions (507 1 keV) + (515 1 keV) to the total counts
(511 10 keV) under the curve [6,7]. While, the S-parame-
ter is sensitive to the open volume defects, the W-parameter
is sensitive to the chemical environment of the annihilation
site [7]. To obtain the S-parameter values corresponding to
the various annihilation states, the experimental S(Ep) curve
is analyzed on the basis of the positron diffusion model using
the VEPFIT program invoking multilayer fitting [8].
GIXRD measurements were carried out on these samples
at a glancing angle of 0.5ꢁ using the STOE high resolution
X-ray diffractometer.
creases, reaches a minimum value at around 4 keV which
characterizes the annihilation of positrons in the Co
overlayer. As the positron energy further increases, the
S-parameter begins to rise and saturates at higher value
beyond about 12 keV which is characteristic of the bulk Si
value. Upon thermal annealing, the following features are
observed: (a) In the temperature range 370–670 K, the
S-parameter decreases due to the annealing of structural
defects in the overlayer. (b) Beyond 670 K, the S-parameter
increases and approaches a value of 0.563 at 1070 K. (c) At
1170 K, the S-parameter drastically increases to ꢀ0.580. It
is also observed that the S vs. Ep curve has considerably
broadened indicating the increase in the thickness of the
overlayer. This implies that the silicide formation has cer-
tainly taken place. The changes in the S-parameter of the
overlayer at various annealing stages suggest that PAS is
also sensitive to the phase transitions in the Co/Si system.
The silicide phases are identified by comparing the exper-
imental S–W coordinates of the overlayer with the calcu-
lated S–W coordinates of the silicide phases, namely,
Co2Si, CoSi and CoSi2. If we assume that the S and W
parameters do not depend on the crystal structure and the
lattice constants of the material, then the S or W value of
a silicide phase is equal to the weighted averages of the S
or W parameter of the constituents in the bi-atomic mate-
rial [9]. For co-existence of more than a single phase, the
effective S or W parameter is the average of S or W param-
eter of the individual silicide phases. Fig. 2 shows the evolu-
tion of the experimental S–W coordinate of the overlayer as
a function of annealing temperature. This is done by taking
an average of the S and W values between 3 and 5 keV
(around the minimum at 4 keV) for the 300–1070 K temper-
ature range and between 1 and 5 keV for 1170 K annealed
sample. As the annealing temperature is increased, the S–
W coordinate decreases up to about 670 K indicating the
3. Results and discussion
Fig. 1 shows the variation of the S-parameter as a func-
tion of positron beam energy, Ep. At low positron energies,
the S-parameter value is high and characterizes the annihi-
lation of positrons at the surface of the Co overlayer. As the
positron energy is increased, the S-parameter at first de-
Fig. 1. S-parameter vs. positron beam energy curves for the Co/Si system.
The mean implantation depth of the positron beam is shown on the top
axis. The vertical dashed line corresponds to the location of the original
interface. Experimental S-parameter values of Co and Si are shown by
horizontal lines. The solid line through the data points is a result of the
VEPFIT analysis.
Fig. 2. S–W correlation plot of the overlayer for various annealing
temperatures. The closed symbols represent the experimental S–W
coordinates and the open symbols represent the calculated S–W coordi-
nates. Experimental values for pure Co and Si are also shown. The solid
line through the data points is a guide to the eye.

2764
S. Abhaya et al. / Surface Science 600 (2006) 2762–2765
annealing of structural defects in the overlayer. The S–W
coordinate at 670 K matches with the S–W coordinate of
the annealed Co state. Beyond 670 K, the S–W coordinate
increases. The increase in the S–W coordinate between 670
and 1070 K is attributed to intermixing of Co and Si across
the interface, prior to silicide formation. At 1070 K, the S–
W coordinate matches with the S–W coordinate of the over-
layer comprising of Co and CoSi suggesting the onset of
silicide phase formation. For 1170 K, the experimental
S–W coordinate is very close to the calculated S–W coordi-
nate of CoSi2 indicating that there is a complete conversion
to CoSi2 at this temperature. Since, the S–W coordinate of
the overlayer matches well with the estimated S-W coordi-
nate of the silicide phase, it is an indication that the silicide
overlayer is free from the vacancy defects.
Fig. 3 shows the GIXRD spectra of Co/Si samples at se-
lected annealing temperatures. As can be seen, only Co lines
are present until about 970 K indicating that no silicide
formation has taken place but intermixing across the inter-
face cannot be ruled out. At 1070 K, there are prominent
Co and CoSi lines and less intense lines due to CoSi2 phase.
At 1170 K, only CoSi2 lines are present. This indicates the
complete conversion to CoSi2 phase. The identification of
the silicide phases at 1070 K and 1170 K compares well with
that identified from the S–W correlation plot.
DEKTAK and RBS estimates. For 1070 K and 1170 K an-
nealed sample, the thickness of the overlayer increased to
95 nm and 400 nm, respectively. The thickness of CoSi2
overlayer at 1170 K is consistent with that calculated for
a given thickness of Co to form CoSi2 [1]. The VEPFIT de-
duced positron diffusion length in the overlayer ranges be-
tween 8 and 15 nm in the temperature range 370–970 K. A
possible reason for observing a smaller positron diffusion
length in the overlayer would be due to the small grain size
of the Co microcrystals. On the other hand, the diffusion
length of positrons in the CoSi2 overlayer turns out to be
ꢀ200 nm. Taking into account, the positron diffusion coef-
ficient [9] to be 2.9 cm2 sÀ1 and the positron lifetime of
CoSi2 as 155 ps [10], the positron diffusion length can be
estimated to be ꢀ210 nm, which compares favorably with
the present experimental value. This indicates that the
CoSi2 overlayer is devoid of vacancy defects. It may also
be pointed out that the estimated S–W coordinate of CoSi2
compares well with the experimental value signifying that
the CoSi2 overlayer is devoid of vacancy defects. Thus, it
is inferred that no vacancy defects are present in the silicide
overlayer as well as at the substrate end of the interface. It
may be pointed out that vacancy defects are created at the
interface consequent to diffusion of either the metal or the
silicon atoms [11,12]. If Si is the dominant diffusing species,
vacancy defects are created at the interface as well as the
substrate region, as observed in the Pd/Si system [13]. On
the other hand, if metal is the dominant diffusing species,
then no vacancy defects are created at the interface or at
the substrate region, which is the case for Ni/Si system
[14]. Our recent Rutherford backscattering spectrome-
try(RBS) studies on the same Co/Si samples [15] have indi-
cated that Co elemental profile gets more broadened as
compared to Si elemental profile consequent to the silicida-
tion, suggesting that Co is the fast diffusing species, which
is consistent with the present observation of absence of va-
cancy defects.
Furthermore, the experimental S vs. Ep curves were ana-
lyzed using the VEPFIT program by assuming a two-layer
model comprising of the overlayer and the substrate region
at various temperature regimes i.e., Co and Si between
370 K and 970 K, CoSi and Si at 1070 K, CoSi2 and Si at
1170 K. From the VEPFIT analysis, the thickness of the
Co overlayer was found to be ꢀ80 20 nm in the temper-
ature range 370–970 K which compares favourably with
4. Conclusions
Co/Si samples subjected to various annealing tempera-
tures have been investigated using positron beam based
Doppler S-parameter and GIXRD measurements. Silicide
formation begins around 1070 K with the overlayer com-
prising of Co, CoSi and small traces of CoSi2. At 1170 K,
there is a complete conversion to the CoSi2 phase. The dif-
ferent silicide phases identified using the S-W correlation
plot and the GIXRD are consistent with each other. Fur-
ther, from the analysis of the positron annihilation data in
terms of the S-W coordinates and the diffusion lengths, it
is inferred that no vacancy defects are present in the inter-
face as well as in the silicide overlayer.
Acknowledgement
We thank Dr. C.S. Sundar for his interest and
encouragement.
Fig. 3. GIXRD spectra of Co/Si samples at selected annealing temper-
atures. Peaks due to Co, CoSi, CoSi2 and Si lines are marked.

S. Abhaya et al. / Surface Science 600 (2006) 2762–2765
[7] P.J. Schultz, K.G. Lynn, Rev. Mod. Phys. 60 (1988) 701;
2765
References
A.P. Knights, G.R. Carlow, M. Zinke-Allmang, P.J. Simpson, Phys.
Rev. B 54 (1996) 13955.
[8] A. Van Veen, H. Schut, J. de Vries, R.A. Hakvoort, M.R. Ijpma,
AIP Conf. Proc. 218 (1990) 171.
[9] A.P. Knights, M.W.G. Ponjee, P.J. Simpson, M. Zinke-Allmang,
G.R. Carlow, Semicond. Sci. Technol. 12 (1997) 173.
[10] B. Barbiellini, P. Genoud, T. Jarlborg, J. Phys.: Condens. Mat. 3
(1991) 7631.
[11] D.S. Wen, P.L. Smith, C.M. Osburn, G.A. Rozgonyi, Appl. Phys.
Lett. 51 (1987) 1182.
[1] S.P. Murarka, Silicides for VLSI Applications, Academic Press, New
York, 1983.
[2] J.M. Poate, K.N. Tu, J.W. Mayer, Thin films – Interdiffusion and
Reaction, Wiley, New York, 1978.
[3] D.S. Wen, P.L. Smith, C.M. Osburn, A. Rozgonyi, J. Electrochem.
Soc. 136 (1989) 466.
[4] W. Lur, J.Y. Cheng, C.H. Chu, M.H. Wang, T.C. Lee, Y.J. Wann,
W.Y. Chao, L.J. Chen, Nucl. Instrum. Meth. B 39 (1989) 297.
[5] G. Amarendra, B. Viswanathan, G. Venugopal Rao, J. Parimala, B.
Purniah, Curr. Sci. 73 (1997) 409.
[12] A.G. Italyantsev, A. Yu. Kuznetsov, Appl. Surf. Sci. 73 (1993) 203.
[13] S. Abhaya, G. Amarendra, G.L.N. Reddy, R. Rajaraman, G.
Venugopal Rao, K.L. Narayanan, J. Phys.: Condens. Mat. 15
(2003) L713.
[6] R. Krause-Rehberg, H.S. Leipner, Positron Annihilation in Semi-
conductors, Springer-Verlag, Berlin, 1998;
K. Saarinen, P. Hautojarvi, C. Corbel, in: M. Stavola (Ed.),
Identification of Defects in Semiconductors, Academic Press, NY,
1998.
[14] S. Abhaya, G. Amarendra, B.K. Panigrahi, K.G.M. Nair, J. Appl.
Phys. 99 (2006) 033512.
[15] S. Abhaya et al., unpublished data.