Positron annihilation studies of silicon-rich SiO2 produced by high dose ion implantation

Article abstract of DOI:10.1063/1.118315

Positron annihilation spectroscopy (PAS) is used to study Si-rich SiO₂ samples prepared by implantation of Si (160 keV) ions at doses in the range 3×10¹⁶-3×10¹⁷ cm^-2 and subsequent thermal annealing at high temperature (up to 1100 °C). Samples implanted at doses higher than 5 × 10¹⁶ cm^-2 and annealed above 1000 °C showed a PAS spectrum with an annihilation peak broader than the unimplanted sample. We discuss how these results are related to the process of silicon precipitation inside SiO₂.

Full text of DOI:10.1063/1.118315

Positron annihilation studies of silicon-rich SiO 2 produced by high dose ion
implantation
G. Ghislotti, B. Nielsen, P. Asoka-Kumar, K. G. Lynn, L. F. Di Mauro, F. Corni, and R. Tonini
Citation: Applied Physics Letters 70, 496 (1997); doi: 10.1063/1.118315
View online: http://dx.doi.org/10.1063/1.118315
View Table of Contents: http://scitation.aip.org/content/aip/journal/apl/70/4?ver=pdfcov
Published by the AIP Publishing
Articles you may be interested in
Positron annihilation studies of the AlOx/SiO2/Si interface in solar cell structures
J. Appl. Phys. 111, 053515 (2012); 10.1063/1.3691895
Identification by photoluminescence and positron annihilation of vacancy and interstitial intrinsic defects in ion-
implanted silicon
J. Appl. Phys. 100, 073501 (2006); 10.1063/1.2354332
Defects in 30 keV Er + -implanted SiO 2 /Si studied by positron annihilation and cathodoluminescence
J. Appl. Phys. 90, 237 (2001); 10.1063/1.1371948
Positron annihilation study of defects in boron implanted silicon
J. Appl. Phys. 87, 7685 (2000); 10.1063/1.373441
Annealing properties of defects during Si-on-insulator fabrication by low-dose oxygen implantation studied by
monoenergetic positron beams
J. Appl. Phys. 87, 1659 (2000); 10.1063/1.372074
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:
150.201.167.219 On: Mon, 15 Dec 2014 03:11:56

Positron annihilation studies of silicon-rich SiO produced by high dose
2
ion implantation
G. Ghislottia) and B. Nielsen
Department of Applied Physics, Brookhaven National Laboratory, Upton, New York 11973
P. Asoka-Kumar and K. G. Lynn
Department of Physics, Brookhaven National Laboratory, Upton, New York 11973
L. F. Di Mauro
Department of Chemistry, Brookhaven National Laboratory, Upton, New York 11973
F. Corni and R. Tonini
Dipartimento di Fisica, Universit a` degli Studi di Modena, Modena, Italy
͑
Received 26 June 1996; accepted for publication 15 November 1996͒
Positron annihilation spectroscopy ͑PAS͒ is used to study Si-rich SiO samples prepared by
2
16
17
Ϫ2
implantation of Si ͑160 keV͒ ions at doses in the range 3ϫ10 –3ϫ10 cm and subsequent
thermal annealing at high temperature ͑up to 1100 °C͒. Samples implanted at doses higher than 5
1
6
Ϫ2
ϫ10 cm and annealed above 1000 °C showed a PAS spectrum with an annihilation peak
broader than the unimplanted sample. We discuss how these results are related to the process of
silicon precipitation inside SiO . © 1997 American Institute of Physics.
2
͓
S0003-6951͑97͒00804-8͔
The study of silicon-rich SiO ͑SiO ) has been the mat-
sidered. In order to discriminate effects related to Si ions
2
x
ter of several papers. Different models have been proposed to
describe this structure: a random-bonding model in which
Si-Si and Si-O bonding are randomly distributed; a random-
implantation from the original Si-SiO interface signal, a sec-
2
ond set of samples, obtained by implanting in fused quartz
Suprasil substrate was considered. Suprasil samples were im-
1
planted at doses from 5ϫ1016 cm to 1.5ϫ10 cm
Ϫ2
17
Ϫ2
.
mixture model for which tetrahedral units are grouped
2
together; finally, a shell model in which silicon clusters are
Positron annihilation spectroscopy measurements were
3
Ϫ7
embedded in stoichiometric SiO . Each of these models
performed in an ultrahigh vacuum chamber (Ͻ10 Torr͒
2
gives a correct picture for a different degree of silicon ex-
cess. Apart from these fundamental questions, Si precipita-
using a variable energy positron beam. In each measurement
10 counts were accumulated. The Doppler broadening of the
6
tion in Si-rich SiO has been used to produce Si nanocrystals
511 keV annihilation line was measured with a HPGe-
detector based gamma spectroscopy system. The broadening
was characterized using the line shape parameter ͑S-
parameter͒, defined as the area of a fixed region (Ϸ1.59 keV
wide͒ in the center of the annihilation peak divided by the
2
͑
Si ) inside a dielectric matrix which can give visible lumi-
nc
4
–6
nescence at room temperature.
In this study we examined annealing behavior of Si-rich
SiO using positron annihilation spectroscopy ͑PAS͒. This
2
8
technique is based on the fact that positrons, when implanted
at a given energy, thermalize and diffuse inside the medium,
and finally annihilate with electrons. Gamma-rays emitted
after the annihilation carry information about the annihilation
site. In a Doppler broadening measurement this information
is extracted by analyzing the broadening of the 511 keV
annihilation peak, due to the non-zero momentum of
total area of the peak. The sharpness of the annihilation
peak is related to the S-parameter, namely sharper peak pro-
duces a higher S-value. In order to compare different S-E
spectra, it is customary to divide measured S-parameter by
that of a reference sample measured under the same experi-
mental conditions. A p-type high resistivity Si sample was
taken as a reference.
8
electrons. The availability of variable energy positron
Suprasil samples were annealed in a vacuum furnace
beams allow non-destructive depth-profiling of materials.8,9
The technique has been extensively used to study the Si-
Ϫ6
(
Ϸ10 Torr͒. Thermal SiO layers on Si samples were an-
2
nealed in situ up to 700 °C with a resistively heated tantalum
SiO system ͑for a review see Ref. 9͒.
Ϫ7
2
foil, at 10 Torr. Above 700 °C, annealing was performed
Samples were prepared by implanting 160 keV 28Siϩ
in N atmosphere. Annealing time was 30 min.
2
ions into 430 nm thick SiO layers thermally grown on a
2
Figure 1 shows the S-parameter values as a function of
the positron implantation energy E ͑S vs E͒ for Suprasil
samples implanted at different fluences. The upper abscissa
represents the mean positron implantation depth z calculated
͑
100͒ oriented p-type Si substrate kept at room temperature
1
6
Ϫ2
during implantation. Fluences ranged from 3ϫ10 cm to
3
1
7
Ϫ2
2
ϫ10 cm , the ion current density was 0.3 ␮A/cm . For a
SiO layer on a silicon substrate, the Si-SiO interface rep-
n
2
2
according to the relation: zϭAE /␳, where, in case of
resents a preferential site for trapping positrons. This gives
rise to a typical Doppler broadening signal with an
3
Ϫ2
Ϫn 9
SiO , ␳ϭ2.33 g/cm , nϭ1.6, and Aϭ4.0 ␮g cm keV
.
2
The S-parameter versus implantation energy curve for unim-
planted sample is also reported. The low S-value observed at
the surface for this sample is related to positrons implanted
at non-zero energy, diffusing back to the surface and annihi-
lating there. When the implantation energy increases,
10
S-parameter lower than that of Si and SiO2 . The effect is
particularly important when low implantation doses are con-
a͒Present address: CORECOM, via Ampere 30, 20131 Milano, Italy.
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:
96 Appl. Phys. Lett. 70 (4), 27 January 1997 0003-6951/97/70(4)/496/3/$10.00 © 1997 American Institute of Physics
50.201.167.219 On: Mon, 15 Dec 2014 03:11:56
4
1

of annealing temperature are reported for samples implanted
at 5ϫ1016 cm and 1ϫ 10 cm , respectively. In the tem-
perature interval from 100 °C to 300 °C the S-parameter in-
creases to values close to 1 ͑the reference silicon S value͒
and it remains constant up to 800 °C. A decrease is observed
only after annealing at 900 °C. Further annealing at 1100 °C
Ϫ2
17
Ϫ2
shows a larger reduction: while for sample implanted at
ϫ1016 cm
Ϫ2
the unimplanted S value is restored, for
5
17
Ϫ2
sample implanted at 1ϫ 10 cm fluence the S-parameter
goes below this limit.
This behavior could be related to defects produced dur-
ing implantation with Si atoms. Different kind of defects can
13
be produced in SiO by implantation: EЈ centers, non-
2
bridging-oxygen hole centers ͑NBOHC͒, and peroxy radi-
cals. In a PAS Doppler broadening spectrum NBOHC give
rise to a low S-value and they are not stable above 300 °C.14
Peroxy radicals are typical of oxygen-rich samples. In a Si-
rich region EЈ defects are more probably produced.1
However, they are positively charged and therefore they can-
not trap positrons efficiently. A high S-value can be related
to positronium (Ps) formation inside voids. In this case ion
implantation should decrease Ps formation by producing
smaller defects and reducing free spaces.17 Therefore S value
in as-implanted samples should decrease as a function of the
dose.
5,16
FIG. 1. Normalized S-parameter vs positron implantation energy for fused
1
6
17
17
Ϫ2
quartz samples implanted at 5ϫ10 , 1 ϫ10 , and 1.5 ϫ10 cm flu-
ences, respectively. The S value for unimplanted sample is also reported.
The insert shows the implantation profile for 160 keV Si ions as calculated
by TRIM code.
S-parameter for the unimplanted sample approaches bulk
value. For an energy of 2 keV ͑mean depth 50 nm͒,
S-parameter is 0.6% lower than bulk value. In order to avoid
any surface effect only regions deeper than 50 nm will be
considered in the discussion of implanted samples.
Assuming that the high S-parameter and its observed
annealing behavior are related to defects, we would expect
that S decreases to the unimplanted value after annealing at
1
100 °C ͑the level reported in Fig. 2 refers to unimplanted
The profile of implanted Si ions as calculated by TRIM
code11 ͑see insert in Fig. 1͒, identifies the region where
sample annealed at 1100 °C͒. This is observed in the sample
implanted at 5ϫ1016 cm but not at 1ϫ 10 cm , where S
decreases below the unimplanted value. The previous argu-
ments exclude that the decrease in S after annealing at
1100 °C is related to defect recovery. The region of interest
corresponds to a Si-rich zone. We propose that the high
S-parameter is related to the presence of Si-atoms in a con-
Ϫ2
17
Ϫ2
SiO is enriched. In the case of implanted samples, compari-
2
son of S vs E curves with TRIM profile shows that
S-parameter is maximum for mean positron implantation
depth close to the range of Si ions. Using VEPFIT code12 we
obtained that, for a positron energy of 5 keV and a diffusion
length of 10 nm ͑typical value for SiO ), 72% of the im-
2
centration higher than that of stoichiometric SiO . The first
2
planted positrons annihilate at a depth ranging from 140 nm
to 330 nm, i.e., in the Si-rich region. Therefore the S value at
annealing stage observed around 300 °C could be due to re-
covery of defects, like EЈ, typical of a Si-rich SiO . It is
2
5
keV can conveniently describe the behavior of the Si-rich
region.
In Fig. 2, S-parameter values for Eϭ5 keV as a function
known that these defects anneal out in this temperature
range.1
8,19
When heated at high temperature, ordering processes can
take place in amorphous SiO , as already observed using
2
PAS.2
1
0,21
However such a phenomena are effective above
200 °C22 or, around 900–1000 °C, for very long annealing
time ͑48 h in Ref. 21͒. The experimental conditions to ob-
serve ordering in SiO2 are therefore different from those
adopted in the present case. Moreover S-value for unim-
planted sample reported in Fig. 2 refers to a sample annealed
under the same conditions than implanted ones. S-value in
this case is higher than for non-annealed sample, whereas
S-parameter is expected to decrease when ordering processes
20
or crystalline phases are present. Our results agree with
data presented by Shimura et al.23 who observed crystalline
phases in as-grown thermal SiO layers, but no changes after
2
annealing for 1 h at 950 °C. Therefore another kind of struc-
tural changes are responsible for S-value decrease in im-
planted samples after annealing at high temperature.
FIG. 2. Normalized S-value at 5 keV ͑corresponding to a mean implantation
depth of 225 nm͒ as a function of the annealing temperature for Suprasil
samples implanted at 5ϫ 1016 cm and 1ϫ10 cm . The unimplanted
Ϫ2
17
Ϫ2
Above 1000 °C SiO dissociates according to the
x
4,25
level is also presented ͑dashed line͒.
reaction:2
SiO →SiϩSiO , therefore the Si excess pre-
x
2
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:
Appl. Phys. Lett., Vol. 70, No. 4, 27 January 1997 Ghislotti et al. 497
50.201.167.219 On: Mon, 15 Dec 2014 03:11:56
1

the Si-SiO interface. This assumption is also consistent with
2
the dose dependence, since for increasing doses the Sinc vol-
ume fraction increases. Hence the ‘‘effective’’ interface and
so the fraction of implanted positrons annihilating there in-
crease.
In conclusion silicon-implanted fused quartz and thermal
SiO layers have been studied using positron annihilation
2
spectroscopy. After annealing at 1100 °C, samples implanted
at 5ϫ1016 cm showed no difference in S-parameter re-
spect to unimplanted samples. Samples implanted at a dose
higher than 5ϫ1016 cm present S-parameter lower than
unimplanted value. This difference is related to silicon pre-
cipitation in SiOx .
Ϫ2
Ϫ2
1
W. Y. Ching, Phys. Rev. B 26, 6610 ͑1982͒.
J. Ni and E. Arnold, Appl. Phys. Lett. 39, 554 ͑1981͒.
P. Bruesch, Th. Stockmeier, F. Stucki, P. A. Buffat, and J. K. N. Linder,
2
3
J. Appl. Phys. 73, 7677 ͑1993͒.
T. Shimizu-Iwayama, K. Fujita, S. Nakao, K. Saitoh, T. Fujita, and N.
Itoh, J. Appl. Phys. 75, 7779 ͑1994͒; T. Shimizu-Iwayama, S. Nakao, and
K. Saitoh, Appl. Phys. Lett. 65, 1814 ͑1994͒.
P. Mutti, G. Ghislotti, L. Meda, E. Grilli, M. Guzzi, L. Zanghieri, R.
FIG. 3. Normalized S vs energy curves for SiO layers on Si implanted at
2
4
1
6
Ϫ2
17
Ϫ2
different doses ͑3ϫ10 cm –3ϫ10 cm ) and subsequently annealed at
000 °C.
1
5
cipitates. Nesbit26 studied Si-rich SiO samples having dif-
Cubeddu, A. Pifferi, P. Taroni, and A. Torricelli, Thin Solid Films 276, 88
2
͑
1996͒.
ferent silicon supersaturation. After annealing at tempera-
tures higher than 900 °C he observed the formation of Sinc
with diameters of 1–10 nm depending on annealing time,
temperature, and degree of supersaturation. In thermal oxide
layers implanted at 2 ϫ 1017 cm and annealed at 1000 °C,
Sinc with dimensions of 3-5 nm were detected using trans-
6
7
T. Komoda, J. Kelly, F. Cristiano, A. Nejim, P. L. F. Hemment, K. P.
Homewood, R. Gwilliam, J. E. Mynard, and B. J. Sealy, Nucl. Instrum.
Methods Phys. Res. B 96, 387 ͑1995͒.
P. Mutti, G. Ghislotti, S. Bertoni, L. Bonoldi, G. F. Cerofolini, L. Meda,
E. Grilli, and M. Guzzi, Appl. Phys. Lett. 66, 851 ͑1995͒.
Ϫ2
8
9
P. J. Schultz and K. G. Lynn, Rev. Mod. Phys. 60, 701 ͑1988͒.
P. Asoka-Kumar, K. G. Lynn, and D. O. Welch, J. Appl. Phys. 76, 4935
6,7
͑
1994͒.
mission electron microscopy. A signature of the presence
10
B. Nielsen, K. G. Lynn, Y. C. Chen, and D. O. Welch, Appl. Phys. Lett.
1, 1022 ͑1987͒.
of Sinc inside SiO is also the observation of a photolumines-
2
5
4
–7
cence ͑PL͒ emission centered at around 780 nm. PL spec-
11
12
J. F. Ziegler, J. P. Biersack, and U. Littmark, The Stopping and Range of
Ions in Solids, Vol. I ͑Pergamon, New York, 1985͒.
tra for samples studied in this work were reported
2
7
A. van Veen, H. Schut, J. de Vries, R. A. Hakvoort, and M. R. Ijpma, in
Positron beams for solids and surfaces, Proceedings of the Fourth Inter-
national Workshop on Slow-Positron Beam Techniques for Solids and
Surfaces, edited by P. J. Schultz, G. R. Massoumi, and P. J. Simpson
͑AIP, New York, 1990͒, p. 171.
elsewhere. The point of interest here is that only for flu-
17
Ϫ2
ences higher than 1ϫ10 cm a PL emission at 780 nm is
observed, thus supporting the idea that Sinc are formed.
The previous arguments lead to relate the observed low
S-value after annealing at 1100 °C with Si precipitation in-
1
1
3
4
D. L. Griscom and E. J. Friebele, Radiat. Eff. 65, 63 ͑1982͒.
E. H. Poindexter, P. J. Caplan, and G. J. Gerardi, in The Physics and
side SiO . This picture is consistent with a shell model for
2
Chemistry of SiO and the Si-SiO Interface, edited by C. R. Helms and B.
3
2
2
SiOx .
E. Deal ͑Plenum, New York, 1988͒, p. 304.
15
Figure 3 shows S vs E curves for SiO layers implanted
H. Hosono and N. Matsunami, Phys. Rev. B 48, 13469 ͑1993͒.
A. Kalnitsky, J. P. Ellul, E. H. Poindexter, P. J. Caplan, R. A. Lux, and A.
R. Boothroyd, J. Appl. Phys. 67, 7359 ͑1990͒.
Mbungu-Tsumbu, D. Segers, M. Dorikens, L. Dorikens-Vanpraet, C.
Laermans, and A. van den Bosch, J. Non-Cryst. Solids 65, 131 ͑1984͒.
P. M. Lenahan and P. V. Dreesendorfer, J. Appl. Phys. 55, 3495 ͑1984͒.
D. L. Griscom, J. Appl. Phys. 58, 2524 ͑1985͒.
2
at fluences ranging from 3ϫ1016 to 3ϫ10 cm after an-
nealing at 1000 °C. Only the Si-rich region is changing as
the dose is increased, and we observe a monotonic decrease
of S for increasing implantation dose.
17
Ϫ2
16
17
18
1
2
9
0
We can discuss now which center can act as annihilation
site for positrons. We will base our discussion on the fact
that: ͑1͒ the signal is related to Sinc formation, ͑2͒ the
S-value is lower than the unimplanted oxide, and ͑3͒ this
reduction has a monotonic dependence on dose ͑see Fig. 2
and Fig. 3͒. The only reported low S-value for the Si-
G. Brauer and G. Boden, Appl. Phys. Lett. 16, 119 ͑1978͒; G. Brauer, G.
Boden, A. Balogh, and A. Andreeff, Appl. Phys. 16, 231 ͑1978͒.
G. Brauer and G. Boden, Diffus. Defect Data Pt. B 53-54, 173 ͑1987͒.
J. Zarzycki and R. Mezard, Phys. Chem. Glasses 3, 163 ͑1962͒.
T. Shimura, I. Takahashi, J. Harada, and M. Umeno, in The Physics and
Chemistry of SiO and the Si-SiO Interface-3, Electrochem. Soc. Proc.,
21
22
23
2
2
Vol. 96-1, edited by H. R. Massoud, E. H. Poindexter, and C. R. Helms
Electrochemical Society, Pennington, NJ, 1996͒, p. 456.
SiO system is observed for positrons trapped at the interface
2
͑
10
and annihilating there. In our case positrons might be
24
D. Dong, E. A. Irene, and D. R. Young, J. Electrochem. Soc. 125, 819
trapped at the interface between Sinc and SiO matrix. Posi-
͑1978͒.
2
25
F. Rochet, G. Dufar, H. Roulet, B. Pelloie, J. Perri e` re, E. Fogarassy, A.
Slaoui, and M. Froment, Phys. Rev. B 37, 6468 ͑1988͒.
L. A. Nesbit, Appl. Phys. Lett. 46, 38 ͑1985͒.
G. Ghislotti, B. Nielsen, P. Asoka-Kumar, K. G. Lynn, A. Gambhir, L. F.
DiMauro, and C. E. Bottani, J. Appl. Phys. 79, 8660 ͑1996͒.
tron diffusion in crystalline silicon is high ͑200 nm͒. Because
of the small size of Si , most of the positrons will diffuse to
26
27
nc
the interface region before annihilation. At this interface pos-
itrons can be trapped with the same mechanism effective for
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:
98 Appl. Phys. Lett., Vol. 70, No. 4, 27 January 1997 Ghislotti et al.
50.201.167.219 On: Mon, 15 Dec 2014 03:11:56
4
1