Hydrostatic pressure dependence of the photoluminescence of Si nanocrystals in SiO2

Article abstract of DOI:10.1063/1.116767

We have measured the hydrostatic pressure dependence of the photoluminescence (PL) of Si nanocrystals in SiO2 layers at room temperature and at pressures up to 50 kbar. The samples were fabricated by ion implantation and subsequent annealing. For the two

Full text of DOI:10.1063/1.116767

Hydrostatic pressure dependence of the photoluminescence of Si nanocrystals in SiO2
Hyeonsik M. Cheong, W. Paul, S. P. Withrow, J. G. Zhu, J. D. Budai, C. W. White, and D. M. Hembree Jr.
Citation: Applied Physics Letters 68, 87 (1996); doi: 10.1063/1.116767
View online: http://dx.doi.org/10.1063/1.116767
View Table of Contents: http://scitation.aip.org/content/aip/journal/apl/68/1?ver=pdfcov
Published by the AIP Publishing
Articles you may be interested in
Size-tunable strain engineering in Ge nanocrystals embedded within SiO2 and Si3N4
Appl. Phys. Lett. 105, 172106 (2014); 10.1063/1.4900942
Optical characterizations of doped silicon nanocrystals grown by co-implantation of Si and dopants in SiO2
J. Appl. Phys. 116, 143505 (2014); 10.1063/1.4898038
Temperature dependent growth morphologies of parahexaphenyl on SiO2 supported exfoliated graphene
J. Vac. Sci. Technol. B 31, 04D114 (2013); 10.1116/1.4813895
Reduction in intermediate layer thickness of CoCrPt – SiO 2 perpendicular recording media by using Ru – SiO 2
J. Appl. Phys. 99, 08E713 (2006); 10.1063/1.2177128
Visible photoluminescence from Ge nanocrystal embedded into a SiO2 matrix fabricated by atmospheric
pressure chemical vapor deposition
Appl. Phys. Lett. 68, 1189 (1996); 10.1063/1.115964
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:
142.157.129.67 On: Tue, 16 Dec 2014 19:07:20

Hydrostatic pressure dependence of the photoluminescence
of Si nanocrystals in SiO2
Hyeonsik M. Cheonga) and W. Paul
Division of Applied Sciences, Harvard University, Cambridge, Massachusetts 02138
S. P. Withrow, J. G. Zhu, J. D. Budai, and C. W. White
Oak Ridge National Laboratory, Oak Ridge, Tennessee 37831
D. M. Hembree, Jr.
Oak Ridge Y-12 Plant, Oak Ridge, Tennessee 37831
͑Received 25 September 1995; accepted 24 October 1995͒
We have measured the hydrostatic pressure dependence of the photoluminescence ͑PL͒ of Si
nanocrystals in SiO2 layers at room temperature and at pressures up to 50 kbar. The samples were
fabricated by ion implantation and subsequent annealing. For the two samples measured, negative
pressure coefficients of Ϫ0.4 and Ϫ0.6 meV/kbar were obtained. These values are not in good
quantitative agreement with an estimate based on the quantum confinement model. Possible
implications of this disagreement are discussed. © 1996 American Institute of Physics.
͓S0003-6951͑96͒03601-5͔
The optical properties of Si-based nanostructures have
been the subject of many studies since the observation of
room-temperature visible PL in porous Si,1 due to their pos-
sible application in integrated optoelectronic devices. De-
spite much effort, the origin of this PL is still unclear. Utili-
zation of porous Si in devices, in any event, is doubtful
because of its structural fragility. Other methods2–6 of pro-
ducing Si nanostructures that luminesce have been reported.
Here, we concentrate on nanocrystals formed by Siϩ implan-
tation into SiO2, followed by thermal annealing.6 The Siϩ
implantation method is particularly attractive both for appli-
cations and for fundamental studies, because SiO2 films con-
taining Si nanoclusters are mechanically and chemically
stable and do not contain hydrogen which may complicate
understanding of the physics of the PL. Visible PL from
Siϩ-implanted SiO2 has been studied by a number of
groups,7–11 and strong red PL from samples with high im-
plantation densities has been observed after annealing at
temperatures higher than 1000 °C.7,9–11 It has been found
that this red PL originates from Si nanocrystals produced by
Si precipitation in SiO2, although there is no agreement on
whether the PL is from quantum confined excitons in the Si
nanocrystals9 or from defect states on the interfaces of the Si
nanocrystals and the SiO2 matrix.7 In this letter, we report the
dependence on hydrostatic pressure of the PL energy for Si
nanocrystals in a SiO2 matrix, and test the quantum confine-
ment model by comparing the experimental results with
theoretical estimates.
ing doses to produce a uniform concentration of 5.3
ϫ1021 cmϪ3 of excess Si in a SiO2 film of ϳ850 nm thick-
ness. The sample was then annealed at 1100 °C for 1 h in a
reducing atmosphere ͑Arϩ4% H2͒ to induce precipitation
and the formation of Si nanocrystals. Sample B was prepared
by the implantation of Si into a SiO2 film of ϳ1600 nm
thickness at a single energy of 400 keV with a dose of 1.5
ϫ1017 cmϪ2, which gives rise to an implanted Si distribu-
tion with a peak maximum at ϳ610 nm and full width at
half-maximum of ϳ280 nm as determined from transport
12
and range of ions in matter ͑TRIM͒ calculations. The peak
Si concentration in this sample is the same as the uniform Si
concentration in sample A. This sample was then annealed
under the same conditions as sample A. All implantations
were performed at room temperature. Figure 1 is a cross-
sectional TEM image of Si nanocrystals in sample A. The
nanocrystals are distributed uniformly throughout the SiO2
layer. The average size is ϳ3 nm in diameter, and the sizes
Our samples were fabricated at Oak Ridge by implanting
Siϩ ions into thermally oxidized Si wafers, followed by ther-
mal annealing. The results for two samples ͑samples A and
B͒ are presented here. Sample A was prepared by implanting
Siϩ ions at six energies between 35 and 500 keV with vary-
FIG. 1. Cross-sectional TEM image of Si nanocrystals in SiO2 ͑sample A͒.
The nanocrystals are distributed uniformly throughout the SiO2 layer. The
average size is ϳ3 nm in diameter, and the sizes are distributed within a
range of 2–4 nm.
a͒
electronic mail: hcheong@nrel.gov
Present address: National Renewable Energy Laboratory, 1617 Cole Blvd.,
Golden, CO 80401.
Appl. Phys. Lett. 68 (1), 1 January 1996
0003-6951/96/68(1)/87/3/$6.00
© 1996 American Institute of Physics
87
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:
142.157.129.67 On: Tue, 16 Dec 2014 19:07:20

FIG. 3. Pressure dependence of the PL peak energy of sample A for two
pressure cycles. The peak shifts to lower energy with pressure at a rate of
Ϫ0.4Ϯ0.2 meV/kbar. No hysteresis is seen.
FIG. 2. PL spectra of sample A at 3.7 and 45.6 kbar. The peak intensities are
normalized for comparison. The sharp peak at ϳ1.78 eV is due to the lumi-
nescence of the ruby chips. The shape and width of the PL peak do not
change much, while the peak energy decreases with pressure.
can be seen that the shape and width of the PL peak remain
the same within experimental uncertainty, while the peak en-
ergy decreases with pressure. The PL energy ͑ϳ1.6 eV͒ com-
pares with an effective-mass calculation,17 which predicts the
PL due to quantum confinement to be at 1.5 eV for 3-nm Si
nanocrystals. Figure 3 shows the pressure dependence of the
PL peak energy of sample A for two pressure cycles. The
peak shifts to lower energy with pressure at a rate of Ϫ0.4
Ϯ0.2 meV/kbar. No hysteresis is seen. Figure 4 shows the
pressure dependence of the PL peak energy of sample B for
two pressure cycles. Unlike the case of sample A, there is a
considerable hysteresis in the first pressure cycle, but not in
the second. The initial hysteresis is thought to be a result of
an irreversible pressure-induced relief of built-in strains in
the nanocrystals, but why this particular sample has built-in
strains is not clear. The result for the second cycle can be
taken as representing the property of the stable state of the
are distributed within a narrow range of 2–4 nm. A detailed
characterization of these samples including their lumines-
cence behavior at atmospheric pressure will be published
elsewhere.13
For high pressure measurements, the samples were
thinned down to ϳ25 ␮m by mechanical lapping, and then
cleaved to dimensions of ϳ150ϫ150 ␮m2. A cleaved piece
of a sample was then loaded into a gasketed diamond anvil
cell ͑DAC͒,14 along with ϳ20 ␮m ruby chips for pressure
calibration. To avoid any possible nonhydrostatic strains on
the sample, helium was used as the pressure medium.15 The
PL measurements were performed at room temperature using
as the excitation source the 514.5 nm line of an argon ion
laser with an intensity of 10–15 W/cm2, focused to a diam-
eter of ϳ200 ␮m. Using the 488 nm line of the argon ion
laser as the excitation source did not result in any appreciable
change in the PL spectra. A magnified image of the pressure
compartment was projected onto a screen in situ to ensure
proper focusing of the laser beam on the sample or on the
ruby chips. The luminescence was dispersed by a Spex
double monochromator and detected with a cooled GaAs
photomultiplier tube. The data were acquired using conven-
tional photon-counting electronics and recorded by a com-
puter. The spectral resolution for the PL measurements was
ϳ0.25 meV, determined by the slit widths of the monochro-
mator. The pressure was calibrated with an accuracy of Ϯ1
kbar using the pressure shift of the ruby R1 luminescence
line. Since SiO2 is more compressible than Si by a factor of
ϳ3.5,16 we can assume that the pressure applied to the
nanocrystals is the same as that applied on the sample. After
each pressure change, the DAC was allowed to stabilize for
about an hour before measurements were performed.
Figure 2 shows the PL spectra of sample A at two pres-
sures. The peak intensities, which fluctuate by ϳ15% be-
tween pressure points, with no correlation to the magnitude
of the pressure applied, are normalized for comparison. It
FIG. 4. Pressure dependence of the PL peak energy of sample B for two
pressure cycles. Unlike the case of sample A ͑Fig. 3͒, there is a considerable
hysteresis in the first pressure cycle, but not in the second. The pressure
coefficient for the peak energy for the second cycle is Ϫ0.6Ϯ0.2 meV/kbar.
88
Appl. Phys. Lett., Vol. 68, No. 1, 1 January 1996
Cheong et al.
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:
142.157.129.67 On: Tue, 16 Dec 2014 19:07:20

sample, and the pressure coefficient of the PL peak energy
for this cycle is Ϫ0.6Ϯ0.2 meV/kbar.
at room temperature and at pressures up to 50 kbar. For the
two samples measured, negative pressure coefficients of
Ϫ0.4 and Ϫ0.6 meV/kbar were obtained. These values are
not in good quantitative agreement with an estimate based on
the quantum confinement model.
The work at Harvard was supported in part by the Na-
tional Science Foundation under Grant No. DMR-91-23829
and by the MRSEC Program of the NSF under Award No.
DMR-9400396. The work at Oak Ridge was sponsored by
the Division of Materials Sciences, U.S. Department of En-
ergy, under Contract DE-AC05-84OR21400 with Lockheed
Martin Energy Systems, Inc.
In the quantum confinement model, the PL energy EPL of
a nanocrystal is determined by the band-gap energy Eg of
bulk Si plus the confinement energy Ec of the electrons and
holes. In a particle-in-a-cubic-box model, Ecϰ(1/L2), where
L is the length scale of the nanocrystal. Therefore, if we take
the pressure dependencies of the effective masses of elec-
trons and holes to be negligible, it can be shown that
dEc /dPϭϩ(2/3)KEc , where K is the volume compress-
ibility of the nanocrystal. A first-principles calculation on the
compressibility of hydrogenated Si clusters shows that a
cluster of 35 Si atoms has a volume compressibility which is
different from that of bulk Si by less than 10%.18 Therefore,
it is reasonable to assume that the compressibility of 3-nm
1 L. T. Canham, Appl. Phys. Lett. 57, 1046 ͑1990͒.
2 T. Kawaguchi and S. Miyazima, Jpn. J. Appl. Phys. 32, L215 ͑1993͒.
3 S. Nozaki, S. Sato, H. Ono, and H. Morisaki, Mater. Res. Soc. Symp.
Proc. 351, 399 ͑1994͒.
nanocrystals
is
the
same
as
that
of
bulk
Si͑1.02ϫ10Ϫ3 kbarϪ1͒.19 If we use EcϭEPL͑ϳ1.6 eV͒–
Eg(ϳ1.1 eV͒Ϸ500 meV, we get dEc /dPϷ0.3 meV/kbar.
Since dEg /dPϭϪ1.5 meV/kbar for silicon, we obtain
dEPL /dPϷϪ1.5ϩ0.3ϭϪ1.2 ͑meV/kbar͒. The above esti-
mate, although simplistic, yields the same results, within
Ϯ0.1 meV/kbar, as a more sophisticated theory,20 which has
been used for Si quantum wires. ͑The above argument ap-
plies to quantum wires as well.͒
4 S. H. Risbud, L.-C. Liu, and J. F. Schackelford, Appl. Phys. Lett. 63, 1648
͑1993͒.
5 L. Brus, J. Phys. Chem. 98, 3577 ͑1994͒.
6 T. Shimizu-Iwayama, M. Ohshima, T. Niimi, S. Nakao, K. Saitoh, T.
Fujita, and N. Itoh, J. Phys. Condens. Matter 5, L375 ͑1993͒.
7 T. Shimizu-Iwayama, S. Nakao, K Saitoh, and N. Itoh, J. Phys. Condens.
Matter 6, L601 ͑1994͒.
8 H. A. Atwater, K. V. Shcheglov, S. S. Wong, K. J. Vahala, R. C. Flagan, M.
L. Brongersma, and A. Polman, Mater. Res. Soc. Symp. Proc. 316, 409
͑1994͒.
Our experimental results, Ϫ0.4 or Ϫ0.6Ϯ0.2 meV/kbar,
do not seem to be in good quantitative agreement with the
above estimated value, although a number of unknown fac-
tors, such as a possible difference in the pressure dependence
of the band gap of the nanocrystals and that of bulk Si, or
different changes in the oscillator strength for different sizes
of nanocrystals upon application of pressure, may explain
this difference still within the quantum confinement model.
Comparison of the present result with parallel measurements
for other semiconductor nanocrystals ͑e.g., Ge21 or GaAs͒ in
SiO2 would provide a more definitive test of the model; a
positive pressure coefficient of the bulk band-gap energy in
these semiconductors should result in a positive pressure co-
efficient for the PL energy of the nanocrystals, if the quan-
tum confinement model is correct.
9 P. Mutti, G. Ghislotti, S. Bertoni, L. Bonoldi, G. F. Cerofolini, L. Meda, E.
Grilli, and M. Gruzzi, Appl. Phys. Lett. 66, 851 ͑1995͒.
10 T. Komoda, J. P. Kelly, A. Nejim, K. P. Homewood, P. L. F. Hemment, and
B. J. Sealy, Mater. Res. Soc. Symp. Proc. 358, 163 ͑1995͒.
11 J. G. Zhu, C. W. White, J. D. Budai, S. P. Withrow, and Y. Chen, Mater.
Res. Soc. Symp. Proc. 358, 175 ͑1995͒.
12 J. P. Biersack and L. G. Haggmark, Nucl. Instrum. Methods 174, 257
͑1980͒.
13 S. P. Withrow, J. G. Zhu, C. W. White, and J. D. Budai ͑unpublished͒.
14 I. F. Silvera and R. J. Wijngaarden, Rev. Sci. Instrum. 56, 121 ͑1985͒.
15 J. H. Burnett, H. M. Cheong, and W. Paul, Rev. Sci. Instrum. 61, 3904
͑1990͒.
16 American Institute of Physics Handbook, 3rd ed., edited by D. E. Gray
͑American Institute of Physics, New York, 1972͒.
17 T. Takagahara and K. Takeda, Phys. Rev. B 46, 15578 ͑1992͒.
18 K. Jackson and E. Kaxiras ͑private communication͒.
19 G. Martinez, in Optical Properties of Solids, edited by M. Balkanski
͑North-Holland, Amsterdam, 1980͒, pp. 181–222.
In conclusion, we measured the hydrostatic pressure de-
pendence of the PL of Si nanocrystals in SiO2 layers, fabri-
cated by ion implantation and subsequent thermal annealing,
20 C.-Y. Yeh, S. B. Zhang, and A. Zunger, Appl. Phys. Lett. 64, 3545 ͑1994͒.
21 C. M. Yang, K. V. Shcheglov, M. L. Brongersma, A. Polman, and H. A.
Atwater, Mater. Res. Soc. Symp. Proc. 358, 181 ͑1995͒.
Appl. Phys. Lett., Vol. 68, No. 1, 1 January 1996
Cheong et al.
89
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:
142.157.129.67 On: Tue, 16 Dec 2014 19:07:20