Dependence of the InAs size distribution on the stacked layer number for vertically stacked InAs/GaAs quantum dots

Abstract

Atomic force microscope (AFM), transmission electron microscopy (TEM), and photoluminescence measurements were carried out to investigate the dependence of the InAs quantum dot size distribution on the stacked layer number for vertically stacked InAs/GaAs quantum dots (QDs) grown on (0 0 1) GaAs substrates. AFM and TEM images showed that the size of the QDs increased with increase in the stacked layer number up to the deposition time of 20 s. However, the size distribution uniformity of the QDs was improved with increase in the stacked layer number when the deposition time and the stacking layer of the InAs QDs gradually decreased. These results can help in an improved understanding of the control of sizes of QDs in InAs/GaAs QD arrays.

Full text

Page 1

Journal of Crystal Growth 241 (2002) 63–68
Dependence of the InAs size distribution on the stacked layer
number for vertically stacked InAs/GaAs quantum dots
H.S. Leea, J.Y. Leea,*, T.W. Kimb, D.C. Choob, M.D. Kimc, S.Y. Seod, J.H. Shind
aDepartment of Materials Science and Engineering, Korea Advanced Institute of Science and Technology, 373-1 Guseong-dong,
Yuseong-ku, Daejeon 305-701, South Korea
bDepartment of Physics, Kwangwoon University, 447-1 Wolgye-dong, Nowoon-ku, Seoul 139-701, South Korea
cOptoelectronics Division, Samsung Electronics, Suwon-City, Kyungki-Do 442-742, Japan
dDepartment of Physics, Korea Advanced Institute of Science and Technology, 373-1 Guseong-dong, Yuseong-ku,
Daejeon 305-701, South Korea
Received 7 February 2002; accepted 18 February 2002
Communicated by M. Schieber
Abstract
Atomic force microscope (AFM), transmission electron microscopy (TEM), and photoluminescence measurements
were carried out to investigate the dependence of the InAs quantum dot size distribution on the stacked layer number
for vertically stacked InAs/GaAs quantum dots (QDs) grown on (001) GaAs substrates. AFM and TEM images
showed that the size of the QDs increased with increase in the stacked layer number up to the deposition time of 20s.
However, the size distribution uniformity of the QDs was improved with increase in the stacked layer number when the
deposition time and the stacking layer of the InAs QDs gradually decreased. These results can help in an improved
understanding of the control of sizes of QDs in InAs/GaAs QD arrays. r 2002 Elsevier Science B.V. All rights
reserved.
PACS: 68.37.Lp; 73.21.La
Keywords: A1. Characterization; A3. Molecular beam epitaxy; B2. Semiconducting III–V materials
1. Introduction
Self-assembled quantum dots (SAQDs) have
been fabricated on various material systems [1–3]
and utilizing various growth techniques by using
the Stranski–Krastanow (S–K) growth mode [4]
since SAQDs have both physical properties of
discrete artificial atom-like energy levels, thermal
stability of states, high gain, d-function-like
density of states, and good optical properties [5],
and potential applications such as quantum dot
lasers [6], optical quantum memories [7,8] and
infrared photodetectors [9,10]. In highly lattice-
mismatched systems, this growth process starts a
two-dimensional mode, and islands are formed
spontaneously remaining the thin wetting layer
under dots beyond a certain critical thickness [11].
Since the SAQDs are formed spontaneously, the
*Corresponding author. Tel.: +82-42-869-4256; fax: +82-
42-869-4276.
E-mail address: jylee@mail.kaist.ac.kr (J.Y. Lee).
0022-0248/02/$-see front matter r 2002 Elsevier Science B.V. All rights reserved.
PII: S 0022 -0248(02 )0 1252-6

Page 2

precise control of the distribution and the size of
the SAQDs are very difficult. Since the size and the
site of the quantum dots (QDs) are randomly
distributed, the full-width
(FWHM) of photoluminescence (PL) spectrum is
broadened [12–14]. It is very important to control
the size of the QDs for improving their optical
efficiency. Recently, it has been found that the
multiple stacking process of the QD layers, which
form the vertical ordering, is useful to control sizes
of the QDs [15,16]. However, the size of the QDs
increases with increasing the stacking layer num-
ber [17]. Therefore, studies concerning the depen-
dence of the InAs size distribution of vertically
stacked InAs/GaAs QDs on the deposition time
are very important.
This paper reports data for the dependence of
the InAs size distribution of vertically stacked
InAs/GaAs QDs grown by using molecular beam
epitaxy (MBE) on the deposition time. Atomic
force microscopy (AFM) and transmission elec-
tron microscopy (TEM) measurements were per-
formed to characterize the surface morphology
and the microstructural properties in InAs QD
arrays inserted into GaAs layers. PL measure-
ments were carried out in order to investigate the
interband transitions in the samples.
athalf-maximum
2. Experiment
The samples used in this work were grown on
semi-insulating (100)-oriented GaAs substrates by
using MBE. The growth rate of the InAs and the
GaAs was 1.15(A/s. The whole growth process was
controlled in situ by reflection high energy electron
diffraction (RHEED). The QD samples studied in
this study consisted of the following structures. A
100nm thick GaAs buffer layer was first grown on
a GaAs substrate at 5301C. Then, the substrate
temperature was lowered to 4101C for the InAs/
GaAs QD array growth. The InAs wetting layer
was deposited followed by a GaAs spacer layer
after a 10s growth interruption. The RHEED
pattern became spotty after the deposition of the
1.8ML InAs layer, indicative of three-dimensional
island formation. After the substrate temperature
was increased to 4301C, a 10nm undoped GaAs
layer was grown. The sample structure for the
multiple stacked InAs/GaAs QDs is the same as
the above-described sample except that the spacer
thickness is 4nm. Five periods of the InAs/GaAs
SAQD arrays were grown at 4101C via the S–K
growth mode. This relatively low temperature
growth reduced the sizes of the QDs. S1 represents
a single layer consisting of QDs, and S2 five
stacked layer QDs without a capping layer. S3, S4
and S5 represent single-stacked, three-stacked and
five-stacked QDs layers with capping layers,
respectively. The deposition times for vertically
stacked InAs/GaAs QDs gradually decreased with
a stacking layer. The depositions of the InAs layers
are summarized in Table 1.
The AFM observations were performed to
measure the size and density of the QDs without
capping layers and carried out to check the vertical
ordering and the size and density of the QDs with
a capping layer, the TEM observations were
performed in a JEM 2000 EX transmission
electron microscope operating at 200kV. The
samples for the cross-section and plan-view TEM
measurements were prepared by cutting and
polishing with a diamond paper to a thickness of
approximately 30mm and then argon-ion milling
at liquid nitrogen temperature to electron trans-
parency. The PL measurements were carried out
using a 0.25m monochromater equipped with a
liquid nitrogen-cooled InGaAs and Si detector.
The excitation source was the 514.5nm line of an
Ar+laser.
Table 1
Deposition times (s) of InAs layers for the samples
InAs layer numberSample
S1S2S3S4S5
1
2
3
4
5
2020
20
20
20
20
2023
14
14
18
16
15
14
14
The last InAs layers of S1 and S2 samples are uncovered. The
last InAs layers of S3, S4, and S5 samples are covered with
GaAs.
H.S. Lee et al. / Journal of Crystal Growth 241 (2002) 63–68
64

Page 3

3. Results and discussion
AFM images for the single and the five-stacked
InAs/GaAs QDs without capping layers are shown
in Figs. 1(a) and (b), respectively. Fig. 1(a) shows a
high density (B9?1010cm?2) of QDs. The
average peak height of the QDs is 3.3nm, and
the average diameter of the QDs about 20nm.
Fig. 1(b) shows that the density of the S2 QDs is
approximately 5?1010cm?2. The average peak
height of the QDs is about 6nm and their average
Fig. 1. AFM images of the top areas of the InAs/GaAs QDs
with (a) single and (b) five-stacked layers samples without
capping layers.
Fig. 2. Cross-sectional micrographs of (a) bright field and (b)
dark field TEM images for the five-stacked InAs/GaAs QDs.
Fig. 3. Cross-sectional micrographs of the InAs/GaAs QDs
with the (a) single, (b) three, and (c) five-stacked layers with
capping layers.
H.S. Lee et al. / Journal of Crystal Growth 241 (2002) 63–68
65

Page 4

diameter of the QDs is approximately 40nm.
These results indicated that the density of the QDs
decreased but the size of the QDs increased with
increasing the stacking layer number.
Cross-sectional micrographs of the S2 sample are
shown in Fig. 2. Fig. 2(a) and (b) are a (004)
bright-field and dark-field TEM image, respectively.
Coherently strained islands are clearly observed for
the S2 sample, which is indicated by the character-
istic dark and the bright contrasts related to an
inhomogeneous lattice deformation in the three-
dimensional structure and to the absence of
dislocations [18]. The strain field surrounding the
islands extends to 5–10nm into the GaAs buffer
and cap layers. Fig. 2 shows that the size of the
upper side of the QDs is larger than that of the
lower side of the QDs, which is also in reasonable
agreement with the results of the AFM images, as
shown in Fig. 1. These results indicated that the size
of the QDs increases with increase in the stacked
layer number up to the deposition time of 20s.
The uniform SAQDs with a high density and a
small size are necessary for device application,
such as QD lasers. One of the possibilities to
achieve the QDs with high density is to stack InAs
QD layers closely separated by a thin GaAs
spacer. However, when the stacked layer number
increases, the size of upper side of the QDs
increases [17]. This result is in reasonable agree-
ment with the AFM results, as shown in Fig. 1. In
Fig. 4. Plan-view micrographs of the InAs/GaAs QDs with the (a) single, (b) three-stacked, and (c) five-stacked layers with capping
layers.
H.S. Lee et al. / Journal of Crystal Growth 241 (2002) 63–68
66

Page 5

order to improve the size uniformity of the stacked
layer QDs, the variation of the size and the optical
properties of the QDs dependent on the deposition
time of the InAs wetting layer have been investi-
gated. The deposition times for the vertically
stacked InAs/GaAs QDs decreased with a stacking
layer number. Figs. 3(a–c) show cross-sectional
micrographs for S3, S4, and S5 samples, respec-
tively. They show vertically aligned quantum
columns. Since the InAs QDs are largely lattice-
mismatched to the GaAs layers, the strain relaxa-
tion can occur. However, there are no dislocations
around the InAs QDs embedded in the GaAs
barriers, as shown in Fig. 3. It is clearly seen that
the size of upper size QDs is almost the same as
that of the lower side QDs.
Figs. 4(a–c) show plan-view micrographs for S3,
S4, and S5, respectively. The apparent diameter of
the islands was determined by measuring the
length of the white and black boundary on the
plan-view micrograph of Fig. 4. Due to the extent
of the strain field around the islands, the apparent
diameter is larger than the actual island diameter.
The distribution of the apparent diameter of the
islands for S3, S4, and S5 is centered between 20
and 30nm. The areal density is approximately
5?1010cm?2, which is in reasonable agreement
with AFM results, as shown in Fig. 1. These
results indicated that the size distribution of the
QDs was improved with increase in the stacked
layer number when the deposition time of the InAs
QDs gradually decreased with stacking layers.
Fig. 5 shows PL spectra at 26K for S3–S5
samples. Strong luminescence peaks correspond-
ing to interband transitions from the ground
electronic subband to the ground heavy-hole band
are observed in all the samples. Only one broad
peak due to the size fluctuation of QDs is observed
for S3 and S5. However, two broad peaks are
observed for S4, which is attributed to the bimodal
distribution in the size of the QDs. The appearance
of two broad peaks is related to the existence of
the large and the small dots, as shown in Fig. 4(b).
The shift in the PL peak position to the low-energy
side or to the high-energy side by changing the
number of the stacking layers is attributed to the
size distribution of QDs. The size of the QD
increases or decreases by increasing the stacking
layer number. While the InAs QDs with smaller
sizes emit at higher energies, the InAs QDs with
larger sizes emit at lower energies. Therefore, the
peak position for the five-stacked QD layers is
located at the lowest energy side. The PL peak
position increases, and the FWHM of the PL peak
at low temperature becomes smaller as the stacked
number increases. The FWHM of the PL peak for
the QDs with five-stacked layers is 84meV, and
this value is the smallest, indicative of the
improvement of the size distribution of the QDs.
Thus, when the deposition time of the InAs
wetting layer is controlled, the narrow size
distribution can be achieved, and the integrated
intensity of the PL peak might be increased.
4. Conclusions
Dependence of the microstructural and the
optical properties on the growth times for verti-
cally stacked InAs/GaAs QDs was investigated.
The size of the QDs increases by increasing the
stacking layer number when the growth time of the
InAs layer is constant. However, the results of
TEM and PL measurements showed that the
narrow size distribution of vertically stacked InAs
QD arrays inserted into GaAs barriers could be
achieved when the growth time of InAs QDs
gradually decreased by increasing the stacking
Fig. 5. The PL spectra at 26K of the InAs/GaAs QDs with the
(a) single, (b) three, and (c) five-stacked layers with capping
layers.
H.S. Lee et al. / Journal of Crystal Growth 241 (2002) 63–68
67