SiGe quantum dots prepared on an ordered mesoporous silica coated Si substrate

Article abstract of DOI:10.1063/1.120085

This letter reports a new way of preparing wafer sized SiGe quantum dots on an ordered mesoporous sol gel silica coated Si. It was found from x-ray diffraction that very good regular layers of mesoscopic sized SiGe quantum dots can be formed in the silica. Initial low temperature photoluminescence measurements show much improved light emission of the buried dots. This technique is a potential low cost method for producing quantum dot arrays.

Full text of DOI:10.1063/1.120085

SiGe quantum dots prepared on an ordered mesoporous silica coated Si substrate
Y. S. Tang, S. Cai, G. Jin, J. Duan, K. L. Wang, H. M. Soyez, and B. S. Dunn
Citation: Applied Physics Letters 71, 2448 (1997); doi: 10.1063/1.120085
View online: http://dx.doi.org/10.1063/1.120085
View Table of Contents: http://scitation.aip.org/content/aip/journal/apl/71/17?ver=pdfcov
Published by the AIP Publishing
Articles you may be interested in
Dramatically enhanced self-assembly of GeSi quantum dots with superior photoluminescence induced by the
substrate misorientation
APL Mat. 2, 022108 (2014); 10.1063/1.4866356
SiGe quantum dot molecules grown on patterned Si (001) substrates
J. Appl. Phys. 104, 044303 (2008); 10.1063/1.2968946
Application of Czochralski-grown SiGe bulk crystal as a substrate for luminescent strained quantum wells
Appl. Phys. Lett. 90, 181914 (2007); 10.1063/1.2735286
Periodic arrays of epitaxial self-assembled SiGe quantum dot molecules grown on patterned Si substrates
J. Appl. Phys. 100, 084312 (2006); 10.1063/1.2358003
Effective compliant substrate for low-dislocation relaxed SiGe growth
Appl. Phys. Lett. 78, 1219 (2001); 10.1063/1.1351520
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: 130.113.76.6
On: Mon, 15 Dec 2014 17:50:00

SiGe quantum dots prepared on an ordered mesoporous silica
coated Si substrate
Y. S. Tang,a) S. Cai, G. Jin, J. Duan, and K. L. Wang
Electrical Engineering Department, University of California at Los Angeles, Los Angeles,
California 90095-1594
H. M. Soyez and B. S. Dunn
Materials Science Department, University of California at Los Angeles, Los Angeles, California 90095
͑
Received 15 July 1997; accepted for publication 29 August 1997͒
This letter reports a new way of preparing wafer sized SiGe quantum dots on an ordered
mesoporous sol gel silica coated Si. It was found from x-ray diffraction that very good regular layers
of mesoscopic sized SiGe quantum dots can be formed in the silica. Initial low temperature
photoluminescence measurements show much improved light emission of the buried dots. This
technique is a potential low cost method for producing quantum dot arrays. © 1997 American
Institute of Physics. ͓S0003-6951͑97͒02643-0͔
Semiconductor quantum dots ͑QDs͒ have attracted
worldwide attention now due to their suitability for quantum
physics study and their potential device applications in both
future microelectronics and optoelectronics industry. Si–
formation of a crystalline SiGe or Ge dot array in the porous
silica media; any over saturated Ge growth on top of the
buried dot array tends to form a larger second dot array,
which has also a very high density and dot size uniformity.
Initial surface morphology of the sample as examined by
atomic force microscope ͑AFM͒ shows the presence of a
very high density dot array. A typical picture is shown in
Fig. 1. Further cross-section scanning electron microscopic
examination of the same sample suggests that some of the
SiGe dots were formed in the pores. The technological ad-
vantages of the QD preparation method are the full Si tech-
nology compatibility, low cost, and applications in a range of
structures and devices, including magnetic dots for magnetic
recording and information storage.
1
SiGe QDs, with their technological compatibility to the Si
industry, are especially important. Recent research in this
field has led to very promising results, such as the efficient
Si–SiGe dry etched QD-based light emitting diodes as re-
2
,3
ported earlier, which could potentially be used in all Si-
based optical interconnects on Si microchips. It was found
that the strong light emission in the dry etched Si–SiGe dots
is accompanied by a simultaneous crystalline lattice shrink-
4
age and distortion, which may have effectively created a
new lattice structure through the combined effect of strain
and dry etching, resulting in an indirect–direct band gap
transition. In this letter, we report a low cost QD fabrication
technique for creating a similar or new lattice distortion in
the SiGe system for light emission application.
The SiGe dot arrays were then studied by both double
axis x-ray diffraction ͑XRD͒ and photoluminescence ͑PL͒.
The XRD was performed on a Crystal Logic powder diffrac-
tometer using the 1.125 kW copper K-␣1 x-ray source
We first prepared an ordered mesoporous silica film5
onto Si substrate by a sol gel process. The film could be spun
on or dip-coated on any substrate such as Si, GaAs, normal
glass, metal sheet, mica, and many others. In our experi-
ments, the silica film was dip-coated on Si at a wafer pulling
speed of about 2 in./min. In the sol preparation, tetraethox-
͑which has a wavelength of ␭ϭ0.154 036 nm͒ and a graphite
ysilane ͑TEOS͒, absolute ethanol, H O, and HCl were mixed
2
and diluted in ethanol in the presence of a surfactant. The
resulted silica material contains regular arrays of uniform
porous tubes with the pore sizes controlled by the choice of
the surfactant species via intercalation of layered silicates. In
this work, we used a mesoporous silica film with pore sizes
of 4.5ϳ5 nm. After coating the film onto a Si substrate, we
grow a thin SiGe or Ge layer onto the ordered mesoporous
silica film at a very slow rate that will allow the SiGe atoms
to diffuse into the opening of the mesoscopic porous tubes.
In our experiments, we used molecular beam epitaxy ͑MBE͒
although the SiGe growth can be done by chemical vapor
deposition, and other techniques. The samples were then heat
treated to distort the lattice necessary for increasing SiGe
based light emission. The post-MBE growth baking helps the
FIG. 1. AFM picture of the top surface of a Si0.5Ge0.5 dot sample after heat
treatment at 500 °C for 90 min.
a͒Electronic mail: ystang@ee.ucla.edu
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: 130.113.76.6
448 Appl. Phys. Lett. 71 (17), 27 October 1997 0003-6951/97/71(17)/2448/3/$10.00 © 1997 American Institute of Physics
On: Mon, 15 Dec 2014 17:50:00
2

FIG. 3. 4.2 K PL spectrum of a Si0.5Ge0.5 dot sample prepared on mesopo-
rous silica coated Si substrate. The sample was post-baked at 500 °C for 45
min.
FIG. 2. Comparison of the XRD spectra from a porous silica coated Si
substrate and a Si0.5Ge0.5 dot sample. The fringes appear around the sub-
strate peak suggests that the dots form regular arrays in the dot sample.
due to strain effect introduced deliberately in the dot prepa-
ration process. The origins of the higher energy peaks from
the buried dots are confirmed by PL measurements carried
out on SiGe QD samples with different Ge contents. It was
found that the energy positions of the higher energy PL
peaks linearly depend on the Ge content in the SiGe, as
shown in the insert of Fig. 3.
monochromator. Assuming there are regular layers of SiGe
QDs formed in the sample, clear fringes related to the intro-
duction of semiconductor atoms into the sample should be
observable.
Shown in Fig. 2 is a comparison of the double axis x-ray
diffraction of a Si Ge QD sample and a sol gel silica
0.5
0.5
To optimize the optical emission intensity for possible
future optoelectronics device applications, a series of post
SiGe growth baking processes has been experimented with.
It was found that the PL intensity increases with increasing
anneal temperature up to 500 °C ͑for a fixed baking time͒,
beyond which the PL intensity drops. The optimum anneal-
ing time for our samples studied here is about 90 min. This is
understandable if one reviews what occurs in the dry etched
Si–SiGe QDs. As we know, the mesoporous silica was pre-
pared by a sol gel process and the increase of annealing
temperature causes the ‘wet’ sol gel silica to shrink its pores
which are partially filled with SiGe after the SiGe deposition.
This drying process of the sol gel silica may alter the SiGe
lattice with a compressed strain. Due to the small size of the
dots, crystal defects are minimized. From the results of small
dry etched Si–SiGe dots, we know that strong light emission
can be realized and the strong light emission is always ac-
companied by a reduced lattice constant and a lattice
coated Si substrate in the same region corresponding to the
buried layers of QDs in the silica media. It can be found that
the Si Ge dot sample has clear observable diffraction sat-
0
.5
0.5
ellites up to the second order centered at around 23.65° off
the ͑004͒ direction of the Si substrate peak, suggesting the
formation of very good regular layers of SiGe dots in the
sample. We suspect that this diffraction pattern comes from
the buried dots inside the mesoporous silica matrix. The
inter-dot separation may be decided by the dimension of the
inter-connected mesoporous silica tubes which are arranged
in such a way that they form layered arrays of pores with the
openings of the pores tilted towards the sample surface. SiGe
atoms then enter the exposed layered pores to form dots.
Other samples with different Ge content in the SiGe alloy are
similar to that we described here for the Si Ge dots.
0.5
0.5
PL has been used to monitor the light emission intensity
of these dots. The measurements were done on a standard
ϩ
cryostat system at 4.2 K. The 488 nm line of an Ar laser
3
was used as the excitation with its power density of
distortion. The same phenomenon could occur in our current
SiGe dots, which may have improved light emission as dem-
onstrated here.
2
4
W/cm . The PL signal was detected by using either a liquid
nitrogen cooled germanium detector or a GaAs photomulti-
plier through a 1.5 m monochromator. The slit sizes were
kept at 150 ␮m for all the measurements.
Shown in Fig. 4 is a typical 4.2 K PL spectrum of an
optimized Ge dot sample prepared in the ordered mesopo-
rous silica matrix. There are two sizes of dots in this sample,
the smaller buried dots in the silica matrix with average pore
sizes of about 4 nm and the bigger sized dots over the top of
the silica film, which is about 100 nm in diameter on average
as seen in AFM. A typical PL of a self-assembled dot sample
with similar dot sizes is also shown on the top panel for
comparison with that of the bigger dots on top of the silica
layer. It is clearly shown that the bigger dots show a broad
Figure 3 shows a typical PL spectrum of a Si Ge QD
0.5
0.5
sample; there are two set of PL features located in two sepa-
rated energy ranges. The low energy features resemble the6
normally observable PL peaks in Si–SiGe heterostructures,
which we believe are from the larger over-grown dots in the
sample; the higher energy peaks at around 1.5–1.6 eV may
be due to the smaller buried dots. The higher energies of the
emissions are partly due to quantum confinement and partly
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: 130.113.76.6
Appl. Phys. Lett., Vol. 71, No. 17, 27 October 1997 Tang et al. 2449
On: Mon, 15 Dec 2014 17:50:00

1
dry etched Si–SiGe or Si–Ge dots as reported before. Un-
like emissions from an ordinary SiGe system, this is similar
to that of a direct band gap material such as GaAs. This
indicates that the higher energy peak is from the buried Ge
dots in the silica.
The shrinking sol gel silica matrix may have the follow-
ing possible effects on the buried dots:
͑
͑
1͒ altering the band structure of the material system due to
strain effect;
2͒ the small dot size could also have strong quantum size
effects although the inter-dots wave function coupling
may also contribute to the improved light emission;
3͒ a large distortion in the symmetry of the squeezed dot
lattice, making our dots into a totally different crystal
structure than Si or Ge and resulting in a direct band gap.
͑
The fact that the phonon replicas, which are usually observ-
able in materials and structures made of the SiGe system, are
not present suggest mechanism ͑3͒ is involved although there
is no direct convincing experimental evidence.
In conclusion, we have proposed a new low cost way of
preparing nano-sized QDs suitable for mass production. Our
initial experimental result on the SiGe system suggests that
SiGe QDs prepared by this method can have much improved
light emission which may be the result of the SiGe lattice or
lattice symmetry change. The improved light emission from
these low cost SiGe dots is ideal for future Si-based opto-
electronics devices.
This work was supported in part by SRC, NSF, and
ARO MURI ͑low power electronics͒ programs.
1
Y. S. Tang, C. M. Sotomayor Torres, T. E. Whall, E. H. C. Parker, H.
FIG. 4. 4.2 K PL spectrum of a Ge dot sample after baking at 500 °C for 90
min. A PL spectrum from a self-assembled MBE Ge dot sample with similar
dot size to the over grown Ge dots is also shown on the top panel for
comparison.
Presting, and H. Kibbel, J. Mater. Sci. 6, 356 ͑1995͒.
Y. S. Tang, W. X. Ni, C. M. Sotomayor Torres, and G. V. Hansson,
Electron. Lett. 31, 1385 ͑1995͒.
Y. S. Tang, C. M. Sotomayor Torres, W.-X. Ni, and G. V. Hansson,
2
3
Superlattices Microstruct. 20, 505 ͑1996͒.
Y. S. Tang, S. Hicks, C. M. Sotomayor Torres, W.-X. Ni, C. D. W.
4
band ͑possibly two unresolved peaks together͒ emission cen-
tered at about 0.78 eV, while an extra peak appears at 1.0223
eV. The latter is about 6.5 times stronger than the broad
band. In addition, there is no phonon replica which can be
observed on the lower energy side of the strong peak at
Wilkinson, and G. V. Hansson, Proc. SPIE 3007, 170 ͑1997͒.
R. Ganguli, Y. Lu, M. T. Anderson, C. J. Brinker, H. Soyez, B. Dunn,
Michael Huang, and Jeff Zink, ‘‘Rapid, continuous formation of supported
mesoporous films from homogeneous sols by a surfactant-templated
mechanisms’’ Nature ͑in press͒.
5
6
M. Wachter, F. Schaffler, H. J. Herzog, K. Thonke, and R. Sauer, Appl.
1.0223 eV, which is almost exactly what we observed in the
Phys. Lett. 63, 376 ͑1993͒.
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: 130.113.76.6
450 Appl. Phys. Lett., Vol. 71, No. 17, 27 October 1997 Tang et al.
On: Mon, 15 Dec 2014 17:50:00
2