Journal of Luminescence 135 (2013) 232238

Contents lists available at SciVerse ScienceDirect

Journal of Luminescence
journal homepage: www.elsevier.com/locate/jlumin

Photoluminescence and structural studies of Tb and Eu implanted

at high temperatures into SiO2 lms
F.L. Bregolin a, U.S. Sias b,n, M. Behar a
a
b

Instituto de Fsica, Universidade Federal do Rio Grande do Sul, Av. Bento Gonc- alves 9500, 91501-970, Porto Alegre-RS, Brazil
Instituto Federal Sul-rio-grandense, Campus Pelotas, Prac- a 20 de Setembro 455, 96015-360, Pelotas-RS, Brazil

a r t i c l e i n f o

a b s t r a c t

Article history:
Received 28 May 2012
Received in revised form
2 October 2012
Accepted 5 October 2012
Available online 16 October 2012

The present work deals with the photoluminescence (PL) emitted from Eu and Tb ions implanted at
room temperature (RT) up to 350 1C in a SiO2 matrix, followed by a further anneal process. The ions were
implanted with energy of 100 keV and a uence of 3  1015 ions/cm2. Further anneals were performed in
atmospheres of N2 or O2 with temperatures ranging from 500 up to 800 1C. PL measurements were
performed at RT and structural measurements were done via transmission electron microscopy (TEM).
In addition, the Rutherford backscattering technique (RBS) was used to investigate the corresponding ion
depth proles. For Tb, the optimal implantation temperature was 200 1C, and the anneal one was of
500 1C. Under these conditions, the PL yield of the sharp band centered at 550 nm was signicatively
higher than the one obtained with RT implants. The PL spectra corresponding to the Eu ions show two
bands, one narrow centered around 650 nm and a second broad one in the bluegreen region.
The implantation temperature plays a small inuence on the PL shape and yield. However, the annealing
atmosphere has a strong inuence on it. Samples annealed in N2 present a broad PL band, ranging from
370 up to 840 nm. On the other hand, the O2 anneal conserves the original as-implanted spectrum, that
is: a broad PL band in the bluegreen region together with sharp PL band in the red one. For both ions,
Tb and Eu, the TEM analyses indicate the formation of nanoclusters in the hot as-implanted samples.
& 2012 Elsevier B.V. All rights reserved.

Keywords:
Earth-rare photoluminescence
Ion implantation
Eu nanoparticles
Tb nanoparticles

1. Introduction
Silicon has been the material of choice for several decades in the
microelectronics industry due to its excellent electrical, chemical
and mechanical properties. With the increasing miniaturization of
the integrated circuits, according to the Moores Law [1], several
difculties are currently hampering their evolution in terms of
processing and transmission speed. Those difculties arise due to
numerous physical factors that become more relevant as the scale
of such circuits becomes closer to atomic dimensions [2]. One
promising alternative is to replace the way that the signals are
transmitted and processed inside the integrated circuits, namely,
the use of photons instead of electrons. However, silicon being an
indirect band gap semiconductor also makes it a poor photoemitter
due to the low probability of radiative recombination.
In the search for the luminescence of silicon-compatible materials,
novel elements were used in order to obtain a more intense emission
with a wider range of wavelengths. Regarding rare earths (RE)
embedded in SiO2 matrixes; the rst successful result related with
PL emission was reported in 1983 by Ennen et al. [3]. In this case,
it was observed that a 1.54 mm emission from Er3 ions was

Corresponding author. Tel.: 55 53 32250586; fax: 55 51 33086510.

E-mail address: uilson.sias@gmail.com (U.S. Sias).

0022-2313/$ - see front matter & 2012 Elsevier B.V. All rights reserved.
http://dx.doi.org/10.1016/j.jlumin.2012.10.010

introduced in a SiO2 matrix by room temperature ion implantation.

In the 90s, the study of the luminescence from RE embedded in Si
compatible matrices started to be intensively investigated. [4]. Several
works have been published involving other REs, like Tb and Eu [47].
In 2007, Prucnal et al. [8] reported an intense electroluminescence
(EL) emission with a strong inuence of the excitation current on the
spectral emission characteristics of the fabricated devices. An extensive review of the subject can be found in Ref. [9].
In all the previous reports where the ion implantation technique
was used, it was performed at room temperature (RT). On the other
hand, very recently it was proved that implantations performed at
high temperatures (hot implants) induce higher PL yields when
compared with the ones performed at RT. In fact, when Si [1014]
and Ge [15] were hot implanted in a SiO2 matrix and further
annealed, PL yields 10 times higher for Si, and three times higher
for Ge were obtained, as compared with the corresponding RT
implantations.
Then, in the present work, we have studied the inuence of the Eu
and Tb hot implants into a SiO2 matrix on the corresponding PL
emission. In addition, the inuence of the annealing temperature and
also the annealing atmosphere were investigated. We have done
structural measurements by using the transmission electron microscopy (TEM) and the Rutherford backscattering technique (RBS) was
used to determine the implanted ion prole at each step of the
experiment.

F.L. Bregolin et al. / Journal of Luminescence 135 (2013) 232238

2. Experimental details
A 100 nm thick SiO2 layer was thermally grown over a o1 0 04
Si substrate. The implantations were performed with a uence of
3  1015 ions/cm2, at an energy of 100 keV (Rp 45 nm, DRp 9 nm),
resulting in peak concentrations of approximately 1 at%. During the
implantations, the samples were kept at a constant temperature,
ranging from RT to 350 1C. Afterwards, the samples were annealed
for 1 h in N2 or O2 atmosphere, with temperatures ranging from 500
to 800 1C.
For the PL measurements, the samples were excited with a
266 nm (4.66 eV) solid state laser, with 7 mW of excitation
power. The samples emission was dispersed by a 30 cm monochromator and then detected by a CCD camera. All the spectra
were corrected taking into account the systems response.
In order to verify the PL emission reproducibility of the
samples and, in particular, to compare the PL behavior of samples
prepared with different parameters we have prepared at least six
independent samples by starting with the same substrate. The
results have shown that statistically the PL shapes and corresponding yields differ in less than 5%.
Rutherford backscattering technique (RBS) and transmission
electron microscopy (TEM) were used in the samples characterization. In the rst case we have used the 1.2 MeV He beam
provided by a 3 MV Tandetron accelerator and for the second one
a 200 kV Jeol transmission electron microscope was used.

3. Results
3.1. Terbium
3.1.1. PL measurements
In Fig. 1(a), a typical PL spectrum of a Tb implanted sample is
displayed. The spectrum presents several emission lines which
correspond to radiative electronic transitions of the Tb ions, as
reported in the literature [16,17]. The shape of the measured Tb
spectra does not change with the implantation or annealing temperatures, neither with the anneal atmosphere, only its intensity is

233

modied. Further, the emission lines maintain their relative intensities. In what follows, we will refer the obtained results only related
to the doublet at 542 nm, which is the most intense PL line.
In Fig. 1(b) the 542 nm PL yield as a function of both the
implantation and annealing temperatures is displayed. An inspection of the gure shows the following features: The implantation
temperature strongly inuences the PL yield. With increasing
implantation temperatures, there is an increase of the PL emission,
being the minimum for the RT and maximum for 350 1C. Concerning the annealing temperature, the maximum is observed at
500 1C regardless of the implantation temperature. However, the
maximum PL yield is observed at a combination of a hot implant
at 200 1C and anneal at 500 1C. Finally, it should be stated that for
temperatures higher than 500 1C, all the samples show a decrease
in the PL yield, independent of the implantation temperature.

3.1.2. RBS results

In Fig. 2 the results of the RBS measurements performed on the
samples implanted at room temperature and at 350 1C, and
further annealed at 800 1C are displayed. As can be observed,
the Tb depth distribution remains unchanged. However, the total
area is 30% smaller as compared with the as implanted prole.
Since no Tb penetration into the Si bulk is observed, it should be
assumed that the Tb out-diffuses through the surface.

3.1.3. TEM results

The TEM imagesFig. 3reveal that the 350 1C implantation
provides sufcient thermal energy for the precipitation of nanoparticles during the implantation process. Further, from the
histograms obtained from the micrographs, the nanoparticles
size distributions of the samples were determinedsee Fig. 3.
For the as-implanted samples at 350 1C, Fig. 3(a), the mean size
diameter is of 2.3 nm, for the RT implants and further annealed at
800 1C, Fig 3(b), it is of 3.2 nm and for the samples implanted at
350 1C and subsequently annealed at 800 1C, Fig. 3(c), it is of
3.9 nm. Fig. 3(d), (e) and (f) are the corresponding histograms
from where the values of the mean size diameters were extracted.

Fig. 1. (a) Typical spectrum of a Tb implanted sample, showing the radiative electronic transitions of the Tb ions. (b) PL intensity of the 542 nm line as function of the
annealing temperature for samples implanted at RT, 200 and 350 1C.

234

F.L. Bregolin et al. / Journal of Luminescence 135 (2013) 232238

3.2.1. PL results
In Fig. 4 the PL spectra of Eu implanted samples at different
temperatures, without further annealing are illustrated. As can be
observed, the bluegreen region of the spectra (400550 nm)
presents a broad PL emission band. On the other hand, in the red
region of the spectra (550800 nm), some emission lines are
resolved being the most intense one situated at around 645 nm.

The presence of narrow emission lines in the red spectral region is

due to the Eu3 , 5D05Fn intra-4f shell transitions, while the large
emission band is due to 5d4f, Eu2 transitions. In addition,
several types of oxygen deciency centers (ODCs) or neutral
oxygen vacancies (NOVs) very likely contribute to this band.
The assignment of the emission lines was done in agreement
with what was reported in the literature [16,17].
An observation of Fig. 4 shows that the PL intensity depends
on the implantation temperature. For samples implanted at RT

Fig. 2. RBS spectra from Tb implanted samples at RT and 350 1C, before and after
an 800 1C, 1 h anneal in a N2 atmosphere.

Fig. 4. PL spectra of the Eu implanted samples at different temperatures.

In addition are shown the electronic transitions responsible for each band.

3.2. Europium

Fig. 3. TEM images for Tb implanted samples: (a) as-implanted at 350 1C, (b) Implanted at RT and further annealed at 800 1C and (c) implanted at 350 1C and subsequently
annealed at 800 1C. Corresponding histograms are shown in gures (d), (e) and (f).

F.L. Bregolin et al. / Journal of Luminescence 135 (2013) 232238

and 200 1C, the intensities are practically the same. On the other
hand, the samples implanted at 350 1C present a 20% increase in
the PL intensity of the lines that are in the red region of the
spectrum, and 45% for the bluegreen PL band, as compared with
the ones implanted at RT.

3.2.2. Annealing results

In Fig. 5 are illustrated the PL spectra of samples implanted at
350 1C and annealed for 1 h in a N2 atmosphere at several
temperatures.
An inspection of Fig. 5 shows that after the thermal treatments, all the spectra present a broad emission band from 370 up
to 850 nm, differing signicantly from the PL emission characteristics of the as-implanted sample. Only for the two lowest
annealing temperatures (500 and 600 1C), a trace of the 645 nm
line is still distinguishable.

Fig. 5. PL spectra from samples implanted at 350 1C and annealed for 1 h in a N2

atmosphere at several temperatures.

235

Further, we have annealed the Eu implanted samples in an O2

atmosphere with temperatures ranging from 500 to 800 1C.
The obtained results are displayed in Fig. 6. The rst remarkable feature is that the spectra of the O2 annealed samples differ
substantially from the N2 annealed ones. Under the present
conditions, the emission spectra have similar spectral characteristics of the as-implanted samples. A second important feature is
that the implantation temperature plays a minor inuence on the
PL yield. Finally, it should be stressed that, with increasing
annealing temperatures, the bluegreen PL emission diminishes,
and is almost completely quenched at 800 1C.
In order to understand the atmosphere inuence on the
observed spectra, the annealed samples were submitted to a
second thermal treatment with a different atmosphere of the
rst one. The samples previously annealed in a N2 atmosphere
were submitted to an additional annealing in O2, also at 700 1C,
and the samples previously annealed at 700 1C in O2 were
subsequently annealed at 700 1C in a N2 atmosphere. In Fig. 7
the corresponding PL results of these experiments are shown.
An inspection of Fig. 7 shows two interesting features: The
emission spectrum of the sample that was rst annealed in N2
and subsequently in O2 changes its shape drastically, becoming
almost identical to the one that was single annealed in an O2
atmosphere. The only difference is that in the double anneal
experiment there is a 25% reduction in the PL intensity. On the
other hand, for a sample rst annealed in O2 and then subsequently in N2, no modication occurs, only a similar reduction in
the PL intensity of 25% is observed. At the end, both PL spectra
(N2 O2 and O2 N2) are practically identical in shape and in
intensity. Then, it can be concluded that the thermal treatment in
O2 atmosphere is the determinant factor that induces a PL shape
similar to the one obtained from the as-implanted samples.
A nal annealing experiment was done with the Eu
as-implanted samples. It was related to the study of the PL
stability, regarding a long time anneal in air (aging). This aging
annealing represents over a year, the oxidation that will suffer an
as-implanted sample being exposed to ambient atmosphere and

Fig. 6. PL spectra of Eu implanted samples at RT and 350 1C: (a) as-implanted, (b) annealed for 1 h in a O2 atmosphere at 500 1C, (c) at 600 1C and (d) at 800 1C.

236

F.L. Bregolin et al. / Journal of Luminescence 135 (2013) 232238

Fig. 7. PL spectra of samples successively annealed in two distinct atmospheres, at (a) 700 1C for 1 h N2 and subsequently with O2 and (b) rst with O2 and then with N2.

An inspection of Fig. 9 indicates the implantation temperature

does not play a signicant inuence on the concentration prole
before the anneals. On the other hand, the 800 1C thermal
treatment induces a strong diffusion toward the surface as well
to the bulk. However, no loss of Eu in the oxide layer was
observed in this last step.
3.2.4. TEM on Eu samples
The TEM imagesFig. 10reveal that the 350 1C implantation
provides sufcient thermal energy for the precipitation of Eu
nanoparticles during the implantation process. Further, the histograms obtained from the micrographs indicate that the mean size
of the nanoparticles are: (a) 4.4 nm for the as-implanted sample
at 350 1C and (b) 9.5 nm for the samples Implanted at 350 1C and
annealed at 800 1C. Before the anneals, no nanoparticle was
observed for the samples implanted at RT.
Fig. 8. PL spectra from samples before and after aging, and after aging and a
subsequent 500 1C 1 h anneal in N2 was performed.

temperature. In the present case, we simulate these conditions

with a sample implanted at 350 1C and subsequently submitted to
a 120 h at 200 1C anneal in air.
In Fig. 8 the results of the present experiment are shown.
At each step of the process PL measurements were taken, showing
the emission spectrum of the as-implanted, as-aged and nally
annealed at 500 1C in N2.
Fig. 8 shows that the aging procedure has a similar effect on
the PL emission shape, as the anneal performed at 800 1C, 1 h, in
an O2 atmosphere. After the 500 1C, N2 thermal treatment, the PL
emission shape does not change signicantly. Only the red
emission band increases its intensity by 50%.

3.2.3. RBS results

In Fig. 9 the RBS spectra for Eu implanted samples at RT and
350 1C before and after the annealing at 800 1C in N2 for 1 h
are shown.

4. Discussion
In the present work, the inuence of the implantation and
annealing temperatures as well the annealing atmosphere on Tb
and Eu implanted samples were investigated.
The PL emission intensity from the samples studied in the
present work as a function of implantation and annealing temperature depend essentially on two conicting factors. On one
hand, the post-implantation anneal favors the nucleation and
growth of the nanoparticles, which is not only unnecessary but
also unwanted. This is because the PL emission from rare earth
oxides is due to electronic transitions of atomic levels of such
ions. On the other hand, non-radiative defects present mainly in
the as-implanted samples, compete in the capture of excitation
photons and even of the emitted photons from the RE luminescent centers, hampering the PL emission. A way to maximize the
PL emission of such systems consists in the passivation of the
non-radiative defects generated by the implantation process and,
at the same time, trying to avoid the growth of the precipitates.
The hot implantation partially eliminates the non-radiative

F.L. Bregolin et al. / Journal of Luminescence 135 (2013) 232238

defects created during the implantation process, which is why the

PL intensity of the as-implanted samples is always higher for the
hot implanted samples as compared to the RT ones.
In the case of Tb, the optimal compromise was obtained for
samples implanted at 200 1C and annealed afterwards at 500 1C.

Fig. 9. RBS spectra from Eu implanted samples at RT and 350 1C before and after
an 800 1C, 1 h in a N2 atmosphere.

237

The annealing atmosphere did not play a signicant role on the

shape or intensity of the PL emission. TEM images revealed that
the mean diameter of the nanoparticles distribution increases
with the implantation temperature and, more signicantly, with
the annealing one. Further, RBS spectra show that samples
annealed at 800 1C had an accentuated out-diffusion of Tb atoms,
causing a loss of 30% of the implanted Tb concentration. However,
the shape of the implantation prole was preserved. These two
factors explain the strong reduction in the PL intensity of the
samples annealed at 800 1C.
For Eu, the situation is more complex. The fact that the
maximum of its ionization energy is between the Eu2 and
Eu3 states favors the presence of both oxidation states in the
SiO2 matrix. Concerning the inuence of the implantation temperature, it was already shown that the samples implanted at
350 1C have higher PL yields as compared with the ones
implanted at RT. The reason is the same as quoted above. A hot
implant partially helps to anneal the defects produced by the
implantation process.
The difference between the spectra of samples annealed in N2
or O2 can be understood based on the following arguments.
As mentioned before, the presence of narrow emission lines in
the red spectral region is due to the Eu3 , 5D05Fn intra-4f shell
transitions, while the large emission band in the bluegreen
region is due to 5d4f, Eu2 transitions. In addition, several types
of oxygen deciency centers (ODC) or neutral oxygen vacancy

Fig. 10. TEM images for Eu implanted samples (a) implanted at 350 1C, and (b) after the 800 1C anneal in N2. The respective histograms are shown in gures (c) and (d).

238

F.L. Bregolin et al. / Journal of Luminescence 135 (2013) 232238

(NOV) defects [18,19] produced by the implantation process can

contribute to this band. Nazarov et al. [19] observed, in an
electroluminescence experiment dealing with Eu implanted in
SiO2, the same broadening effect after a high temperature N2
anneal. They claimed that the high temperature anneal results in
an enhanced clustering of EuOx complexes with a strong generation of ODC around the clusters. These ODC contribute to the
broadening of the PL band.
On the other hand, for Eu implanted samples annealed in O2,
the oxygen atmosphere provides an excess of it, which is supplied
to the system and consequently favors the Eu3 oxidation state.
Then, after the anneal process it can be observed that the
as-implanted PL shape is conserved. With increasing annealing
temperatures (up to 800 1C), the bluegreen emission band,
described in the literature [18] as being due to Eu2 ions, NOV
and ODC practically vanishes. At this temperature, all the defects
are annealed and what is left can be ascribed to Eu2 transitions.
Further, the oxygen passivates both radiative and non-radiative
defects and plays an important part in inhibiting the Eu diffusion
in the oxide matrix, since the diffusivity of EuxOy-like molecules is
signicantly lower than that of a single Eu ion.
Another evidence of the above described behavior is shown in
the experiment where samples were submitted to two successive
thermal treatments in different atmospheres. The nitrogen and
subsequently oxygen annealed samples present a signicant
change in the emission spectra. This change is due to the
passivation of the radiative and non-radiative defects, as well as
favoring of the Eu3 state. Both features are due to the available
excess of oxygen. The samples submitted to the aging treatment
also show a similar behavior to the ones annealed at high
temperatures in oxygen atmosphere.

5. Summary
In the present work we have investigated the inuence of the
implantation temperature, temperature and atmosphere of
annealing on the PL emission of Tb and Eu implanted into a
SiO2 matrix. The microstructural characterization was done with
TEM images and the resulting histograms. The concentration
depth proles of the implanted ions were obtained at each step
of the experiments by using the RBS technique.
In the case of Tb as-implanted samples, it was observed that
the hot implantation gives the highest PL intensity as compared
with the ones implanted at RT. The hot implantation technique
induces concurrently, the nucleation of nanoparticles and the
passivation of defects inside the oxide matrix. It was also shown
that after the hot implants, nanostructures are already present in

the SiO2 lm. The thermal treatment at 500 1C gives the highest
PL intensity for both implanted ions, regardless of the implantation temperature. TEM images show that the mean size distribution of the nanoparticles increases with increasing annealing
temperatures. On the other hand, the annealing atmosphere does
not have any inuence on the Tb PL results.
For the Eu-implanted samples, we also have observed that the
implantation temperature increases the PL yield as compared
with the RT implants. However, the PL enhancement is smaller in
comparison with the Tb implanted samples. At variance with
what was observed for Tb, the annealing atmosphere plays a
major role in the Eu PL results. In fact samples annealed in N2
presented a broad PL band, ranging from 370 up to 840 nm. On
the other hand, the O2 anneal conserves the original shape of the
as-implanted spectra, that is: a broad PL band in the bluegreen
region together with sharp PL band in the red region. This feature
is due to the strong passivation effect of the oxygen atmosphere
together with its inuence in favoring the Eu3 state.
References
[1] P.S. Peercy, Nature 406 (6799) (2000) 1023.
[2] P.J. Pauzauskie, P. Yang, Mater. Today 9 (10) (2006) 36.
[3] H. Ennen, J. Schneider, G. Pomrenke, A. Axmann, Appl. Phys. Lett. 43 (10)
(1983) 943.
[4] S. Wang, H. Amekura, A. Eckau, R. Carius, C. Buchal, Nucl. Instrum. Methods
Phys. Res. B 148 (1-4) (1999) 481.
[5] M. Yoshihara, A. Sekiya, T. Morita, K. Ishii, S. Shimoto, S. Sakai, Y. Ohki, J. Phys.
D-Appl. Phys. 30 (13) (1997) 1908.
[6] C. Buchal, Nucl. Instrum. Methods Phys. Res. B 166 (2000) 743.
[7] J.M. Sun, W. Skorupa, T. Dekorsy, M. Helm, L. Rebohle, T. Gebel, J. Appl. Phys.
97 (12) (2005) 7.
[8] S. Prucnal, J.M. Sun, W. Skorupa, M. Helm, Appl. Phys. Lett. 90 (18) (2007) 3.
[9] L. Rebohle, W. Skorupa, Rare-Earth Implanted MOS devices for Silicon
Photonics: Microstructural, Electrical and Optoelectronic Properties, 1st ed.,
Springer-Verlag, Berlin, 2010.
[10] U.S. Sias, E.C. Moreira, E. Ribeiro, H. Boudinov, L. Amaral, M. Behar, J. Appl.
Phys. 95 (9) (2004) 5053.
[11] U.S. Sias, L. Amaral, M. Behar, H. Boudinov, E.C. Moreira, E. Ribeiro, J. Appl.
Phys. 98 (3) (2005) 034312.
[12] U.S. Sias, L. Amaral, M. Behar, H. Boudinov, E.C. Moreira, Nucl. Instrum.
Methods Phys. Res. B 242 (12) (2006) 109.
[13] U.S. Sias, M. Behar, H. Boudinov, E.C. Moreira, Nucl. Instrum. Methods Phys.
Res. B 257 (2007) 51.
[14] U.S. Sias, M. Behar, H. Boudinov, E.C. Moreira, J. Appl. Phys. 102 (4) (2007)
043513.
[15] F.L. Bregolin, M. Behar, U.S. Sias, E.C. Moreira, Nucl. Instrum. Methods Phys.
Res. B 267 (2009) 1321.
[16] G.H. Dieke, Spectra and Energy Levels of Rare Earth Ions in Crystals,
Interscience, New York, 1968.
[17] G. Liu, B. Jacquier, Spectroscopic Properties of Rare Earths in Optical
Materials, Tsinghua University Press and Springer, Berlin, 2005.
[18] S. Cotton, Lanthanides and Actinides, Macmillan, Basingtoke, 1991.
[19] A.N. Nazarov, S.I. Tiagulsky, P. Tiagulski, V.S. Lysenko, L. Rebohle, J. Lehmann,
S. Prucnal, M. Voelskow, W. Skorupa, J. Appl. Phys. 107 (2010) 123112.