Effects of interdiffusion on the luminescence of InGaAs/GaAs quantum dots
R. Leon, Yong Kim, C. Jagadish, M. Gal, J. Zou, and D. J. H. Cockayne
Citation: Applied Physics Letters 69, 1888 (1996); doi: 10.1063/1.117467
View online: http://dx.doi.org/10.1063/1.117467
View Table of Contents: http://scitation.aip.org/content/aip/journal/apl/69/13?ver=pdfcov
Published by the AIP Publishing
Articles you may be interested in
Thermally activated carrier transfer and luminescence line shape in self‐organized InAs quantum dots
Appl. Phys. Lett. 69, 3354 (1996); 10.1063/1.117304
Photoluminescence studies of InGaAs/InAlAs strained double quantum wells
J. Appl. Phys. 80, 5348 (1996); 10.1063/1.363474
Photoinduced structures in the exciton luminescence spectrum of InGaAs/GaAs quantum well heterostructures
J. Appl. Phys. 80, 3011 (1996); 10.1063/1.363160
Emission from discrete levels in self‐formed InGaAs/GaAs quantum dots by electric carrier injection: Influence of
phonon bottleneck
Appl. Phys. Lett. 68, 3013 (1996); 10.1063/1.116681
Observation of temperature‐insensitive emission wavelength in GaInAs strained multiple‐quantum‐wire
heterostructures
Appl. Phys. Lett. 68, 2386 (1996); 10.1063/1.116142
Reuse of AIP Publishing content is subject to the terms at: https://publishing.aip.org/authors/rights-and-permissions. Download to IP: 130.102.82.188 On: Thu, 29 Sep 2016
03:50:11
Effects of interdiffusion on the luminescence of InGaAs/GaAs
quantum dots
R. Leon,
a)
Yong Kim, and C. Jagadish
Department of Electronic Materials Engineering, Research School of Physical Sciences and Engineering,
Australian National University, Canberra, ACT 0200, Australia
M. Gal
School of Physics, University of New South Wales, P.O. Box 1, Kensington, New South Wales 2033,
Australia
J. Zou and D. J. H. Cockayne
Electron Microscope Unit & Australian Key Centre for Microscopy and Microanalysis, The University
of Sydney, New South Wales 2006, Australia
~Received 29 April 1996; accepted for publication 10 July 1996!
Large energy shifts in the luminescence emission from strained InGaAs quantum dots are observed
as a result of postgrowth annealing and also when raising the upper cladding layer growth
temperatures. These blueshifts occur concurrently with narrowing ~from 61 to 24 meV! of the full
width at half-maxima for the emission from the quantum dot ensemble. These energy shifts can be
explained by interdiffusion or intermixing of the interfaces rather than strain effects due to
variations in capping layer thickness. Temperature behavior of the luminescence in annealed and
nonannealed samples indicates a change in the shape and depth of the quantum dot confining
potential. Quenching of the wetting layer luminescence after interdiffusion is also observed.
© 1996 American Institute of Physics. @S0003-6951~96!00139-8#
Blueshifts in quantum wells ~QW! can be induced by
intermixing
1
assisted by dielectric capping,
2,3
impurity
diffusion,
4
ion implantation,
5
and simple thermal treatments.
6
Interdiffusion of heterointerfaces is expected to play a major
role in structures of reduced dimensionality where the area of
the interface is increased due to island formation, and where
the overall dimensions are small ~15–50 nm diam!. Blue-
shifts in luminescence could be adjusted in situ ~in the
growth chamber, during or after growth! offering a range of
tunability that might be desirable to incorporate into devices.
Achieving narrower luminescence lines for a large en-
semble of quantum dots ~QDs! is a major goal for both future
device applications and fundamental studies of the physics of
zero-dimensional ~0D! structures. Size uniformity correlates
with inhomogeneous broadening of the photoluminescence
~PL! emission.
7
It has been predicted
8
that the 0D laser prop-
erties of low current threshold and higher quantum efficien-
cies can only be of benefit if size uniformity is achieved.
The presence of a so called ‘‘wetting layer’’ @effectively
a very thin quantum well ~1–4 ML! connecting the islands#,
is an inevitable result of the Stranski–Krastanow
9
growth
mode upon reaching a strain-defined critical thickness. This
wetting layer has been shown to have its own distinct lumi-
nescence emission,
10
often appearing as a shoulder on the
brighter PL emission peak for the quantum dot ensemble. Its
presence might be responsible for some nonideal 0D behav-
ior reported in these structures
10,11
as well as constituting a
technical barrier in the development of 0D lasers.
Results presented here show that blueshifts in the PL
emission from a large number of quantum dots can be ob-
tained reproducibly. These blueshifts occur as a result of
thermally induced interdiffusion and are concurrent with nar-
rowing of the full width at half-maxima ~FWHM! of the PL
emission from the QDs. Furthermore, quenching of the wet-
ting layer luminescence is observed, suggesting the possibil-
ity of using interdiffusion to achieve more nearly ideal three-
dimensional confinement in semiconductor quantum dots.
These structures were grown by metalorganic chemical
vapor deposition using a horizontal reactor cell operating at
76 Torr. A specially designed laminar flow cell allows large
areas of uniform growth. Partial pressures for (CH
3
)
3
Ga and
(CH
3
)
3
In were 5.36310
2 6
and 5.183 10
2 6
, respectively.
AsH
3
was used for the group V source and the V/III ratio
was 351. The hydrogen flow rate was 17.5 standard liters per
minute. The flow of (CH
3
)
3
In was monitored and controlled
by an EPISON ultrasonic sensor.
After growth of a GaAs buffer layer at 650 °C on semi-
insulating ~100! GaAs substrates, quantum dots in the form
of nanometer size InGaAs islands, were grown by depositing
4.5 ML ~nominally! of In
0.49
Ga
0.51
As at 550 °C. The tem-
perature was raised to the chosen GaAs upper cladding
growth temperature while growing the GaAs capping layer,
or the GaAs capping layer was grown at the same growth
temperature as the islands. Except in one case, the capping
layer thicknesses were nominally 100 nm and a similar layer
containing InGaAs islands was grown on the surface. The
surface was kept in an inert atmosphere and scanning probe
microscopy ~Nanoscope III with etched SiN tips! was used to
verify island formation and obtain structural information on
average size and areal density.
Post-growth annealing was done in argon using a rapid
thermal annealer at temperatures of 850–950 °C for 30 s.
Low-temperature ~12 K! photoluminescence spectra were
obtained using the 488 nm line of an argon ion laser and
dispersed using a 75 cm spectrometer. The signal was col-
lected using a Si detector and lock-in techniques.
Plan-view transmission electron microscopy ~TEM!
specimens were prepared by chemical etching from the sub-
a!
Electronic mail: rp1109@rsphysse.anu.edu.au
1888 Appl. Phys. Lett. 69 (13), 23 September 1996 0003-6951/96/69(13)/1888/3/$10.00 © 1996 American Institute of Physics
Reuse of AIP Publishing content is subject to the terms at: https://publishing.aip.org/authors/rights-and-permissions. Download to IP: 130.102.82.188 On: Thu, 29 Sep 2016
03:50:11
strate side using H
2
SO
4
:H
2
O
2
:H
2
O58:1:1 aftera1min
etch in H
2
SO
4
:H
2
O
2
:H
2
O(51:8:500) to remove the top
surface QDs. TEM specimens were investigated in a Philips
EM 430 operating at 300 keV.
Results from two growth experiments are presented: in
one, the same QD structure was annealed at different tem-
peratures; in the other, the structures were grown separately
but identical conditions were used except for the temperature
of the 100 nm upper cladding layer. In one case, the thick-
ness of the upper cladding was changed to eliminate possible
in lapido strain effects
12
and both layers were grown at the
same temperature as the QDs.
The preannealing average diameters and areal concentra-
tions were 43 nm and 13 10
10
/cm
2
for the QDs producing
the luminescence in Fig. 2 and 33 nm to 93 10
9
/cm
2
for Fig.
3. Figure 1 shows a scanning probe image of the surface
morphology for one of the samples showing the nanometer
size InGaAs islands. Figure 2 shows dark field plan-view
TEM images of both unannealed and annealed samples with
QDs.
All structures showed very bright QD luminescence.
Figure 3 shows that a large blueshift can be observed with a
corresponding narrowing of the peak with progressively
higher annealing temperatures. Emission from GaAs ~free
and impurity related excitonic transitions! and wetting layer
luminescence are also shown. Figure 4 shows PL spectra for
quantum dot samples grown under identical conditions but
with varying upper cladding growth temperature.
FIG. 1. Morphology of one of the samples used in this study before anneal-
ing or capping layer growth as imaged using scanning probe microscopy.
The width of the scan is 750 nm.
FIG. 2. Plan-view TEM 220 dark-field images taken from an ~a! unannealed
and ~b! annealed at 950 °C quantum dot samples, showing that quantum dots
are still present after annealing.
FIG. 3. Low-temperature ~12 K! photoluminescence spectra showing emis-
sion from quantum dots in as-grown and annealed samples. The smaller
peak at 1.5 eV is due to free and impurity bound excitonic transitions in the
GaAs buffer layer and substrate. Peak A is from an InGaAs/GaAs quantum
dot sample where the quantum dots were grown at 550 °C and the GaAs
buffer and cladding layers were grown at 650 °C. This sample was then
annealed for 30 s at 850 °C ~peak B!, 900 °C ~peak C!, and 950 °C ~peak D!.
The maximum blueshift observed in the sample annealed at the highest
temperature is 140 meV, and the FWHM for the inhomogeneously broad-
ened peak changes from 61 to 24 meV.
FIG. 4. Low-temperature PL for QD structures with different capping layer
growth temperatures. This is equivalent to in situ annealing at a temperature
of 675 °C for 12 min for the sample with emission peak labeled ‘‘3’’ and
800 °C for also 12 min for the peak labeled ‘‘4.’’ Peak ‘‘1’’ shows emission
from a sample where the upper cladding was grown at 550 °C. Another
sample also grown with an upper cladding growth temperature of 550 °C but
much thinner ~20 nm! produces PL emission peak ‘‘2.’’ The broader emis-
sion from peak ‘‘1’’ is due to islands with less uniformity in size. The
average diameters for the islands before annealing was equivalent in all
these samples. The magnified signal at higher energies for ‘‘1’’ and ‘‘2,’’ is
attributed to the wetting layer. Wetting layer luminescence is not observed
for samples ‘‘3’’ and ‘‘4.’’
1889Appl. Phys. Lett., Vol. 69, No. 13, 23 September 1996 Leon
et al.
Reuse of AIP Publishing content is subject to the terms at: https://publishing.aip.org/authors/rights-and-permissions. Download to IP: 130.102.82.188 On: Thu, 29 Sep 2016
03:50:11
The normalized PL signal as a function of temperature
for the unannealed sample in Fig. 1 ~peak A! and for the
sample after annealing at 950 °C ~peak D! are displayed in
Fig. 5, showing that the temperature dependence of the lu-
minescence changes significantly after annealing.
The observed blueshifts in Fig. 3 indicate a trend with
higher temperature annealing. These shifts are a remarkable
change, given the short annealing times involved in the ex-
periment. Initial enhancement of interdiffusion and non-
Fickian behavior in strained systems
6
is most likely at play.
Interdiffusion in systems where strain is as large as in the
samples studied in this work ~3.5%! has not been systemati-
cally studied for quantum wells. It is expected, however, that
the transient component of the diffusivity and deviation from
Fick’s law would be of even greater importance than in the
strained systems studied by S. W. Ryu et al.,
6
where
In
0.2
Ga
0.8
As was used ~;1.4% lattice mismatch!. The large
strain present in In
0.49
Ga
0.51
As QDs might result in a greater
transient component of the diffusion causing a large effect
even for short annealing times.
Narrower FWHM with higher annealing temperatures
could be explained by a homogenization of sizes for the
individual InGaAs islands. Narrower FWHM luminescence
peak for QD ensembles has been correlated with narrower
size distribution.
7
These changes could also be partly ac-
counted for by a smaller expected variation in emission in
dots of different sizes with a shallower confining potential.
Plan-view TEM of annealed and unannealed quantum dot
samples shows that the annealing process does not destroy
the QDs, but a weaker strain contrast results from the ther-
mal treatment. This observation is consistent with interdiffu-
sion of the InGaAs/GaAs interface. A more detailed TEM
study of the effect of intermixing on QD size, QD size varia-
tions, and changes in QD shape is in progress.
13
The fact that blueshifts are obtained upon increasing the
cladding layer growth temperature indicates that the ob-
served shifts in emission energies are due to interdiffusion
rather than strain effects from the capping layer. This result
might have the implication that most of the recent PL studies
of self-organized QD emission arise from quantum dot
samples that do not have abrupt or square confining poten-
tials. The fact that the PL did not shift for caps grown at the
same temperature as the dots when the cap thickness was
changed from 20 to 100 nm indicates that strain effects are
unimportant beyond capping layer thickness above the reach
of the strain field ~around 20–30 nm above QDs!.
To first order, the quenching of the luminescence as a
function of temperature can be modeled by thermal emission
of the carriers out of the quantum dots. Previous measure-
ments including quantum dot systems with different confin-
ing potentials show similar behavior.
14
An Arrhenius plot of
log normalized luminescence intensity shows a thermally ac-
tivated nonradiative recombination mechanism; the slopes of
the straight portions of Fig. 5 give an activation energy re-
lated to the depth of the quantum dot confining potential.
15,16
The lower temperature quenching in the annealed samples as
well as the lower activation energy extracted from the slope
in the Arrhenius plot ~280–120 meV change! can be inter-
preted as a change both in the depth and in the shape of the
confining potential caused by interdiffusion during anneal-
ing.
In summary, large blueshifts result from thermal anneal-
ing of strained InGaAs/GaAs quantum dots. The FWHM
from the quantum dot ensemble PL also becomes narrower,
and wetting layer luminescence indistinct. Similar blueshifts
are obtained with high upper cladding growth temperatures,
indicating that most samples are already blueshifted, and that
interdiffusion rather than strain effects are responsible for
these shifts.
The authors would like to thank A. Clark and T. Thom-
son for technical assistance, and A. Sikorski for his help with
TEM samples preparation. Financial support from the Aus-
tralian Research Council is gratefully acknowledged.
1
For recent review papers on quantum well intermixing: J. H. Marsh, Semi-
cond. Sci. Technol. 8, 1136 ~1993!; D. G. Deppe and N. Holonyak, Jr., J.
Appl. Phys. 64, R93 ~1988!.
2
E. S. Koteles, B. Elman, P. Melman, J. Y. Chi, and C. A. Armiento, Opt.
Quantum Electron. 23, S779 ~1991!.
3
R. M. Kolbas, Y. L. Hwang, T. Zhang, M. Prairie, K. Y. Hsieh, and U. K.
Mishra, Opt. Quantum Electron. 23, S805 ~1991!.
4
W. D. Laidig, N. Holonyak, Jr., M. D. Camras, K. Hess, J. J. Coleman, P.
D. Dapkus, and J. Bardeen, Appl. Phys. Lett. 38, 776 ~1981!.
5
H. H. Tan, J. S. Williams, C. Jagadish, P. T. Burke, and M. Gal, Appl.
Phys. Lett. 68, 2401 ~1996!.
6
S. W. Ryu, In Kim, B. D. Choe, and W. G. Jeong, Appl. Phys. Lett. 67,
1417 ~1995!, and references therein.
7
R. Leon, S. Fafard, D. Leonard, J. L. Merz, and P. M. Petroff, Appl. Phys.
Lett. 67, 521 ~1995!.
8
C. Weisbuch and J. Nagle, in Science and Engineering of One- and Zero-
Dimensional Semiconductors, edited by S. P. Beaumont and C. M. Soto-
mayor Torres ~Plenum, New York, 1989!.
9
I. N. Stranski and Von L. Krastanow, Akad. Wiss. Lit. Mainz Math.
Naturwiss. K1. 146, 797 ~1939!.
10
S. Fafard, D. Leonard, J. L. Merz, and P. M. Petroff, Appl. Phys. Lett. 65,
1388 ~1994!.
11
S. Fafard, R. Leon, D. Leonard, J. L. Merz, and P. M. Petroff, Phys. Rev.
B 52, 5752 ~1995!.
12
M. E. Pistol, N. Carlsson, C. Persson, W. Seifert, and L. Samuelson, Appl.
Phys. Lett. 67, 1438 ~1995!.
13
J. Zou, D. J. H. Cockayne, and R. Leon ~unpublished!.
14
S. Raymond, S. Fafard, S. Charbonneau, R. Leon, D. Leonard, P. M.
Petroff, and J. L. Merz, Phys. Rev. B 52, 238 ~1995!.
15
J. D. Lambkin, D. J. Dunstan, K. P. Homewood, L. K. Howard, and M. T.
Emeny, Appl. Phys. Lett. 57, 1986 ~1990!;
16
G. Bacher, H. Schweizer, J. Kovac, A. Forchel, H. Nickel, W. Schlapp,
and R. Losch, Phys. Rev. B 43, 9312 ~1991!.
FIG. 5. Normalized PL intensity as a function of 1/k
B
T for the unannealed
and annealed samples with PL emissions labeled ‘‘A’’ and ‘‘D’’ in Fig. 3.
1890 Appl. Phys. Lett., Vol. 69, No. 13, 23 September 1996 Leon
et al.
Reuse of AIP Publishing content is subject to the terms at: https://publishing.aip.org/authors/rights-and-permissions. Download to IP: 130.102.82.188 On: Thu, 29 Sep 2016
03:50:11
