4726
Appl. Phys. Lett., Vol. 81, No. 25, 16 December 2002
Chen et al.
FIG. 2. Depth profiles of Si nanocrystals obtained from the spectral fittings
by using the dielectric function of bulk crystalline Si or the dielectric func-
tion of Si nanocrystal calculated with either the bond contraction or phe-
nomenological models for ⌬Eg .
FIG. 3. Comparison of the result of SIMS with the depth profile obtained
from the spectral fitting by using the dielectric function of Si nanocrystal
calculated with the bond contraction model for ⌬Eg .
mine depth profiles of Si nanocrystals in SiO films based on
spectroscopic ellipsometry. In the ellipsometry analysis, the
2
the spectral fitting discussed previously is now carried out by
searching for the nanocrystal size and one set of
SiO film with a depth distribution of Si nanocrystals is di-
2
(
, ,... ) such that the error function ͑F͒ is a minimum.
1 2 m
vided into m sublayers with equal thickness ͑a better depth
resolution for a larger m͒, and each sublayer is characterized
by its nanocrystal concentration. As the dielectric function of
Figures 1͑b͒ and 1͑c͒ show the best fittings with the band
contraction model ͑the second model͒ and the phenomeno-
logical model ͑the first model͒ for ⌬E , respectively. As can
g
SiO is well known, the effective dielectric function of each
2
be seen clearly in Fig. 1, these fittings, which have consid-
ered the nanosize effect on the band gap, are superior to the
fitting with the dielectric function of bulk crystalline Si, in-
dicating that the nanosize effect plays a role in the dielectric
function. The depth profiles of Si nanocrystal concentration
obtained from the fittings based on the two models are
shown in Fig. 2, and they are also compared in this figure
with the depth profile obtained from the fitting using the
dielectric function of bulk crystalline Si. As can be seen in
Fig. 2, the three depth profiles appear to be very similar ͑only
some small difference exist in the peak regions of the pro-
files͒. The reason for this is that, as indicated by our calcu-
lations based on the two models, the difference of dielectric
functions between the bulk crystalline Si and the Si nano-
crystal is not large, as the size of the Si nanocrystals is not
very small. The fitting based on the bond contraction model
yields a nanocrystal size of about 5 nm, which agrees with
the result reported in Ref. 8 for similar nanocrystal formation
conditions, while the fitting based on the phenomenological
model yields a size of about 9 nm. For such nanocrystal
sizes, the bond contraction model ͑and the phenomenological
model͒ gives a decrease of 9% ͑and 16%͒ in the real part and
a decrease of 5% ͑and 9%͒ in the imaginary part of the
dielectric function at the wavelength of 400 nm. Therefore,
although the fittings can be improved by taking into account
the small reductions in the dielectric function due to the size
effect ͑as shown in Fig. 1͒, the reductions do not have a very
significant impact on the depth profiles. The depth profiles
agree with that from the SIMS measurement. Figure 3 shows
the comparison of the depth profile obtained from the spec-
tral fitting based on the bond contraction model with that
from the secondary ion mass spectroscopy measurement. The
good agreement indicates that the above approach for depth
profiling is reliable.
sublayer can be calculated with the effective medium ap-
proximation if the dielectric function of Si nanocrystal is also
given. In the spectral fittings, the dielectric function of Si
nanocrystal is calculated based on two different models ͑i.e.,
the bond contraction model and the phenomenological
model͒ for the band-gap expansion due to the nanocrystal
size reduction. The fittings yield the nanocrystal depth profile
and the nanocrystal size as well. The depth profiles from the
two models are very similar, and they are in good agreement
with the SIMS measurement.
1
L. Brus, in Light Emission in Silicon: From Physics to Devices, Semicon-
ductors & Semimetals Vol. 49, edited by D. Lockwood ͑Academic, New
York, 1998͒, pp. 303–328.
S. Tiwari, F. Rana, H. Hanafi, A. Hartstein, E. F. Crabb e´ , and K. Chan,
Appl. Phys. Lett. 68, 1377 ͑1996͒.
2
3
4
E. Kapetanakis, P. Normand, and D. Tsoukalas, Appl. Phys. Lett. 77, 3450
͑
2000͒.
E. A. Boer, M. L. Brongersma, and H. A. Atwater, Appl. Phys. Lett. 79,
91 ͑2001͒.
7
5
6
S.-H. Choi and R. G. Elliman, Appl. Phys. Lett. 75, 968 ͑1999͒.
M. L. Brongersma, A. Polman, K. S. Min, and H. A. Atwater, J. Appl.
Phys. 86, 759 ͑1999͒.
S. Guha, J. Appl. Phys. 84, 5210 ͑1998͒.
B. Garrido, M. L o´ pez, O. Gonz a´ lez, A. P e´ rez-Rodriguez, J. R. Morante,
7
8
and C. Bonafos, Appl. Phys. Lett. 77, 3143 ͑2000͒.
S. Guha, S. B. Qadri, R. G. Musket, M. A. Wall, and T. Shimizu-Iwayama,
J. Appl. Phys. 88, 3954 ͑2000͒.
E. A. Irene, in In Situ Real-Time Characterization of Thin Films, edited by
O. Auciello and A. R. Krauss ͑Wiley, New York, 2001͒, pp. 57–103.
R. M. A. Azzam and N. M. Bashara, Ellipsometry and Polarized Light
9
10
11
͑
North-Holland, Amsterdam, 1977͒.
1
1
1
2
3
4
C. C. Katsidis, D. I. Siapkas, A. K. Robinson, and P. L. F. Hemment, J.
Electrochem. Soc. 148, G704 ͑2001͒.
C. Q. Sun, X. W. Sun, B. K. Tay, S. P. Lau, H. T. Huang, and S. Li, J.
Phys. D 34, 2359 ͑2001͒.
V. Ranjan, M. Kapoor, and V. A. Singh, J. Phys.: Condens. Matter 14,
6
647 ͑2002͒.
15
C. Q. Sun, H. Q. Gong, P. Hing, and H. T. Ye, Surf. Rev. Lett. 6, 171
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP: 193.0.65.67
In conclusion, we have developed an approach to deter-
͑1999͒.
On: Mon, 22 Dec 2014 13:43:37