Inhibited carrier transfer in ensembles of isolated quantum dots
C. Lobo
Department of Electronic Materials Engineering, Research School of Physical Sciences and Engineering, Australian National University,
Canberra ACT 0200, Australia
R. Leon
Jet Propulsion Laboratory, California Institute of Technology, Pasadena, California 91109
S. Marcinkevic
˘
ius
Department of Physics-Optics, Royal Institute of Technology, S-100 44 Stockholm, Sweden
W. Yang and P. C. Sercel
Department of Physics and Materials Science Institute, University of Oregon, Eugene, Oregon
X. Z. Liao, J. Zou, and D. J. H. Cockayne
Australian Key Centre for Microscopy and Microanalysis, Electron Microscopy Unit, University of Sydney, Sydney NSW 2006, Australia
共Received 7 June 1999; revised manuscript received 16 August 1999兲
We report significant differences in the temperature-dependent and time-resolved photoluminescence 共PL兲
from low and high surface density In
x
Ga
1⫺ x
As/GaAs quantum dots 共QD’s兲. QD’s in high densities are found
to exhibit an Arrhenius dependence of the PL intensity, while low-density 共isolated兲 QD’s display more
complex temperature-dependent behavior. The PL temperature dependence of high density QD samples is
attributed to carrier thermal emission and recapture into neighboring QD’s. Conversely, in low density QD
samples, thermal transfer of carriers between neighboring QD’s plays no significant role in the PL temperature
dependence. The efficiency of carrier transfer into isolated dots is found to be limited by the rate of carrier
transport in the In
x
Ga
1⫺ x
As wetting layer. These interpretations are consistent with time-resolved PL mea-
surements of carrier transfer times in low and high density QD’s. 关S0163-1829共99兲04748-7兴
I. INTRODUCTION
Time-resolved and temperature-dependent photolumines-
cence 共PL兲 measurements of quantum dot 共QD兲 ensembles
have helped to clarify the processes of energy relaxation and
energy transfer in multiple and coupled quantum dot
systems.
1–5
However, the influence of dot size and density
on the luminescence energies and linewidths from QD en-
sembles is still not well understood, and the mechanisms
involved in carrier relaxation and PL quenching seem
strongly dependent on the material system, excitation condi-
tions, and method of QD formation 共self-assembled or strain-
induced兲. It has recently been reported that the PL emission
energies, inhomogeneous linewidths, intersublevel energies,
and excited state relaxation times of ensembles of
In
x
Ga
1⫺ x
As/GaAs QD’s are strongly influenced by the QD
density.
6
Here we have investigated the luminescence emis-
sion from low density and high density In
x
Ga
1⫺ x
As/GaAs
QD’s by temperature-dependent and time-resolved PL. We
find that the temperature dependence of the PL emission is
determined primarily by the efficiency of carrier thermal
transfer into the QD’s. We also report differences in the de-
gree of linewidth broadening of the PL emission from these
low and high density QD’s as a function of temperature.
These differences result from the reduced influence of ther-
mally activated carrier transfer and strain interaction as the
average dot separation increases.
II. EXPERIMENTAL METHODS
QD structures of nominal composition In
0.6
Ga
0.4
As were
grown on slightly misoriented, semi-insulating GaAs共100兲
substrates by metal organic chemical vapor deposition in a
horizontal reactor cell operating at 76 Torr. The structures
were grown under identical conditions, except for the AsH
3
partial pressure, which was varied in order to obtain widely
differing densities of similar-sized In
x
Ga
1⫺ x
As islands.
7
QD
samples used for photoluminescence measurements were
capped with 100 nm GaAs. Full details of the growth condi-
tions are given in Ref. 7.
Average dot sizes in the capped samples used for PL mea-
surements were determined by plan-view transmission elec-
tron microscopy 共TEM兲 using a Phillips EM430 operating at
300 keV. Island concentrations were determined both by
TEM of capped samples and by atomic force microscopy
共AFM兲 of uncapped QD samples grown under identical con-
ditions to the capped samples used for PL. Temperature de-
pendent PL measurements were undertaken with above-
band-gap excitation using a Ti:sapphire laser emitting at 804
nm pumped by an argon ion laser (⫽ 488 nm), or directly
using an argon ion laser. The signal was dispersed by a 0.25
m single grating monochromator and collected using a
liquid-N
2
-cooled germanium detector. QD samples were
mounted with indium to prevent sample heating. The average
excitation power was 1 W/cm
2
.
Carrier transfer times into the QD’s were measured by
PHYSICAL REVIEW B 15 DECEMBER 1999-IIVOLUME 60, NUMBER 24
PRB 60
0163-1829/99/60共24兲/16647共5兲/$15.00 16 647 ©1999 The American Physical Society
time-resolved photoluminescence in the temperature range
80–300 K. A self-mode-locking Ti:sapphire laser 共80 fs, 95
MHz, 800 nm兲 was used for excitation, and an upconversion
technique with a temporal resolution of 150 fs used for sig-
nal detection. Excitation powers varied in the range 0.01 to
10 mW. The average excitation intensity for the temperature-
dependent measurements was 10 W/cm
⫺ 2
, which corre-
sponds to approximately 2⫻ 10
11
electron-hole pairs per
square centimeter when spot size and reflections are taken
into account.
III. TEMPERATURE-DEPENDENT
PHOTOLUMINESCENCE
Figure 1 shows plan-view TEM and AFM images of rep-
resentative high and low density QD samples. The low-
density QD samples A and B 关shown in Figs. 1共a兲 and 1共c兲,
respectively兴 have nominal miscut angles of 0.25° and 0.75°
off the 共100兲 plane. These samples have dot densities of ap-
proximately 3.5⫻ 10
8
cm
⫺ 2
and 7⫻ 10
8
cm
⫺ 2
. Average
共edge to edge兲 QD separations are 280 nm in sample A and
190 nm in sample B. The high density QD sample shown in
Fig. 1共b兲 and 1共d兲 is nominally on axis
关
(100)⫾ 0.05°
兴
and
has an average dot density of 2.5⫻ 10
10
cm
⫺ 2
and an average
QD separation of 10 nm. The average diameter of dots in all
three capped samples is (25⫾ 5) nm.
Low-temperature PL spectra from high and low density
QD’s are displayed in Fig. 2. Emission from the ground state
and two excited states is observed from the low-density
sample even under conditions of low excitation. This may be
due to hindered intersublevel relaxation.
8–10
In contrast, the
high density QD’s show only ground state emission under
continuous-wave 共CW兲 excitation. The line shape of the PL
emission from high-density QD’s is unchanged with increas-
ing excitation power.
6
The high-density dots also exhibit a
blueshift of the PL emission energy and a broadening of the
ground-state inhomogeneous linewidth with respect to the
low-density QD’s at the same low temperature. These differ-
ences have been ascribed to progressive strain deformation
of the QD confining potentials, which effectively result in
shallower confinement as the dot density increases.
6
The in-
creased degree of strain interaction between neighboring
high-density QD’s, together with the randomness of these
interactions, account for the observed differences in the CW
PL. From here on these low- and high-density QD’s will
therefore be referred to as ‘‘isolated’’ and ‘‘interacting,’’ re-
spectively.
A. Interacting QD’s
The normalized PL intensity from interacting QD’s is
plotted as a function of temperature in Fig. 3. These QD’s
FIG. 1. Plan-view TEM images of capped samples used for PL
measurements, and AFM images of uncapped samples grown under
identical conditions. 共a兲 Plan-view TEM image of isolated QD’s
with density 3.5⫻ 10
8
cm
⫺ 2
共sample A). 共b兲 Plan-view TEM image
of interacting QD’s with density 2.5⫻ 10
10
cm
⫺ 2
. The scale applies
to both TEM images. 共c兲 2⫻ 2
m
2
AFM deflection image of iso-
lated QD’s with density 7⫻ 10
8
cm
⫺ 2
共sample B). 共d兲 1⫻ 1
m
2
AFM deflection image of interacting QD sample shown in 共b兲. The
average size of QD’s in all three capped samples is (25⫾ 5) nm.
FIG. 2. Low-temperature 共70 K兲 PL spectra of isolated and in-
teracting QD’s. The spectrum from the isolated QD’s 共sample A)is
modeled as the sum of Gaussians with peak energies E
0
⫽ 1.045 eV, E
1
⫽ 1.094 eV, E
2
⫽ 1.139 eV, and ⌫⫽ 21.8 meV.
FIG. 3. Normalized integrated PL intensity as a function of tem-
perature for several interacting and isolated QD samples. Isolated
QD samples A and B are represented by filled gray and black tri-
angles, respectively; PL data from a third sample is plotted with
unfilled triangles. The integrated intensity from the In
x
Ga
1⫺ x
As WL
in sample A is plotted with gray diamonds. PL data from the inter-
acting QD’s is plotted with black squares. The inset is an Arrhenius
fit to the PL intensity from the interacting QD’s as a function of
temperature.
16 648 PRB 60
C. LOBO et al.
exhibit the Arrhenius behavior typically observed for quan-
tum wells
11,12
共see inset, Fig. 3兲. The activation energy E
A
extracted from these Arrhenius plots approximates the en-
ergy difference between the QD emission and that of the
wetting layer 共WL兲 or barrier. Emission from the WL in
interacting QD samples is absent except at very low tempera-
tures (⬍ 20 K) or under high-excitation densities. Reported
observations of a large increase in QD luminescence when
the excitation energy exceeds the WL emission energy,
13
and
of an increase in the WL/QD intensity ratio with excitation
power
10
indicate that carriers are generated in the WL 共and
GaAs barrier兲 and subsequently transferred to the QD’s.
Hence, the lack of WL emission in CW PL indicates rapid
carrier transfer from the WL to the high-density QD’s.
The temperature dependencies of the PL redshift and in-
homogenously broadened linewidth 共full width at half maxi-
mum, FWHM兲 of the QD emission are plotted in Fig. 4. The
QD PL redshift is significantly greater than that of the
In
x
Ga
1⫺ x
As free exciton emission
14
at temperatures above
100 K. The FWHM of the QD PL emission first decreases
and then increases over an 8 meV range with increasing tem-
perature. Similar dependencies of the energy shift and
FWHM of the QD PL peak on temperature have been re-
ported previously
15–17
and attributed to the effects of thermal
activation transfer and tunneling transfer between neighbor-
ing QD’s 共see Fig. 5兲. At low temperatures, there is no sig-
nificant thermal emission of carriers out of the QDs. Hence,
the PL redshift of the QD peak is equivalent to that of the
In
x
Ga
1⫺ x
As band gap and the FWHM remains constant.
Above ⬇100 K, carrier thermal emission out of the smaller
QD’s in the ensemble and recapture by neighboring QD’s
with deeper confining potentials becomes significant. This
behavior explains the simultaneous redshifting and narrow-
ing of the PL peak. At higher temperatures 共above ⬇150 K)
the FWHM increases and the rate of redshift decreases as
thermal emission of carriers out of the larger QD’s 共and re-
capture by all dots兲 in the ensemble becomes significant.
Filling of the excited states of the larger dots in the ensemble
may also contribute to the broadening of the FWHM ob-
served at high temperature. Tunneling of carriers between
neighboring QD’s may play a role, although this process is
independent of temperature and would lead to a reduction of
the FWHM across the whole temperature range. Attempts to
model the observed behavior by accounting only for carrier
thermal transfer between neighboring QD’s have been only
partially successful.
15,16
In particular the FWHM at high T is
often greater than that at low T, an observation that cannot be
explained by a simple thermionic emission model. A suc-
cessful model may require inclusion of the effects of excited
state filling and varying strain interactions between neighbor-
ing QD’s.
B. Isolated QD’s
PL emission from the ground and excited states of the
isolated QD’s can be modeled by Gaussians with a constant
inhomogeneous broadening factor (⌫⫽ 1/2FWHM) and
level spacings of 45–50 meV.
10
The integrated PL intensity,
energy shift, and FWHM extracted from these fits to the PL
spectra are plotted as a function of temperature in Figs. 3 and
4. The integrated intensity of the luminescence emission
from the ground and excited states increases from 20–90 K,
and subsequently decreases up to 210 K 共see Fig. 3兲. The
total PL intensity at 90 K is approximately twice the low-
temperature value. At low temperature, the wetting layer
emission is more intense than the QD emission. The inte-
FIG. 4. 共a兲 Redshift of the PL peak energy as a function of
temperature for the interacting and isolated 共sample A) QD’s plot-
ted with that of the In
x
Ga
1⫺ x
As free exciton emission. 共b兲 Inhomo-
geneous linewidth broadening 共FWHM兲 as a function of tempera-
ture for interacting and isolated 共sample A) QD’s.
FIG. 5. Schematic of thermally activated carrier transfer be-
tween neighboring quantum dots in high-density QD structures.
PRB 60
16 649INHIBITED CARRIER TRANSFER IN ENSEMBLES OF...
grated intensity of the WL emission as a function of tem-
perature is also plotted in Fig. 3. This plot shows that the
wetting layer emission obeys the Arrhenius dependence ob-
served for quantum wells.
11,12
Wetting layer emission spectra obtained at a number of
temperatures are shown in Fig. 6. The line shape of the broad
WL emission changes with increasing temperature. The in-
tensity of the low-energy tail increases relative to the high-
energy tail, indicating the existence of potential fluctuations
in the WL. Such fluctuations may be caused by indium seg-
regation and enrichment in the quantum dots,
18
which would
produce gallium-rich regions in the WL. Variations in strain
induced by the QD’s would also produce regions of lateral
confinement in the WL.
10
The observed change in the line
shape of the WL emission with increasing temperature indi-
cates that some carriers thermally emitted from regions of
the WL with shallow confining potentials are retrapped and
recombine in regions of deeper confinement. This behavior is
analogous to retrapping of carriers thermally emitted from
shallow wells by adjacent deeper wells in multiple quantum
well samples.
19
It has been established that carrier transfer to these
In
x
Ga
1⫺ x
As/GaAs quantum dots occurs via the WL.
10,13
The
WL behaves as a reservoir of carriers, available to the QD’s
provided the two-dimensional 共2D兲 diffusion length in the
WL is long enough for capture into the QD’s to occur prior
to recombination in the WL. In high-density QD’s, the ab-
sence of a WL emission in CW PL indicates that the average
dot-dot separation is shorter than the carrier diffusion length
in the WL.
10
The carrier mobility in the WL would be ex-
pected to play a greater role in the rate of carrier transfer to
the QD’s as the average dot-dot separation increases.
We attribute the increase in PL intensity from isolated
QD’s 共Fig. 3兲 to carrier transfer mechanisms in the WL. At
elevated temperatures, carriers trapped at potential fluctua-
tions in the WL may acquire sufficient thermal energy to
overcome these barriers. Such an increase in carrier thermal
transfer within the WL would result in an increased rate of
carrier capture by the QD’s and produce the observed in-
crease in QD PL intensity. The rate of carrier transfer to the
QD’s is also limited by the rate of lateral transport in the
In
x
Ga
1⫺ x
As WL, which for photoexcited carriers is gov-
erned by hole diffusion. Hole mobility in In
x
Ga
1⫺ x
As in-
creases with temperature up to ⬇100 K.
20
This increase in
hole mobility may increase the rate of carrier capture into the
QD’s and therefore contribute to the observed increase in
QD PL intensity. At temperatures greater than 100 K, ther-
mal emission of carriers from the QD’s becomes dominant.
The QD PL intensity therefore follows an approximately ex-
ponential decrease up to room temperature.
The PL redshift of the emission from the ground state and
first excited state of the isolated QD’s is plotted as a function
of temperature in Fig. 4. The redshift of both emissions fol-
lows that of the In
x
Ga
1⫺ x
As free exciton emission with in-
creasing temperature. The inhomogeneously broadened line-
width of the isolated QD’s shows an approximately linear
decrease from 45 to 41 meV over the temperature range 10–
210 K. We attribute these behaviors to thermal emission of
carriers from the smaller QD’s in the ensemble without sub-
sequent recapture by larger QD’s. Thus carrier transfer be-
tween neighboring QD’s has no significant effect on the PL
temperature dependence.
IV. TIME-RESOLVED PHOTOLUMINESCENCE
Carrier transfer in isolated and interacting quantum dot
samples was also investigated by time-resolved photolumi-
nescence. The rise times of the luminescence emission from
the isolated and interacting QD’s have been studied as a
function of excitation intensity and temperature 共see Fig. 7兲.
These rise times were measured at the ground state PL peak
energies, and account for carrier transport, capture and relax-
ation in the QD’s. Fast carrier transfer into the interacting
QD’s is confirmed by the short PL rise times shown in Fig.
7. The rise times for interacting QD’s show a slight decrease
FIG. 6. Luminescence emission from the wetting layer in iso-
lated QD sample A at several temperatures.
FIG. 7. PL rise times as a function of 共a兲 excitation power and
共b兲 temperature for interacting and isolated 共sample B) QD’s.
16 650 PRB 60
C. LOBO et al.
with increasing excitation power and display no significant
dependence on temperature. The magnitudes and behavior of
the PL rise times for interacting QD’s are very similar to
those observed for quantum wells of similar composition.
21
For the isolated QD’s, the PL rise times decrease from 15
to 5 ps with increasing excitation power from 0.01 to 10
mW, and increasing temperature from 80 to 300 K. The ex-
citation power dependence may be explained by a reduced
importance of diffusion limited transport at high carrier den-
sities. As the excitation power increases, a greater number of
carriers are generated in the vicinity of the isolated QD’s.
Hence, the PL rise time of the isolated QD’s approaches that
of the interacting QD’s at very high-excitation densities. Po-
tential barriers around the isolated QD’s induced by band
bending
22
would also be reduced at high-carrier densities due
to screening of the internal electric field.
23
The decrease of
the PL rise time with temperature is consistent with an in-
creased rate of carrier transfer due to carriers having greater
thermal energy to overcome these potential barriers around
the QD’s or in the WL.
The temperature and excitation power dependence of the
rise times of the PL emission from isolated QD’s confirm the
interpretation of the PL temperature dependence in terms of
carrier transfer mechanisms in the WL. Further studies will
be undertaken on InAs QD’s, in which the effects of indium
segregation and enrichment in the QD’s and associated com-
positional fluctuations in the WL would be minimized.
24
Differences in the PL temperature dependence of
In
x
Ga
1⫺ x
As/GaAs and InAs/GaAs QD’s may clarify the ef-
fects of temperature-dependent hole mobility and potential
fluctuations in the WL on carrier transfer to isolated quantum
dots.
V. CONCLUSION
In conclusion, we have conducted a study of the
temperature-dependent and time-resolved PL from interact-
ing and isolated In
x
Ga
1⫺ x
As quantum dots. We find that for
high-density QD samples, the temperature dependence of the
PL emission is determined by carrier thermal emission and
recapture into neighboring QD’s. In low-density QD’s the
efficiency of carrier transfer into the dots is limited by the
rate of carrier transport in the In
x
Ga
1⫺ x
As wetting layer.
Thermal transfer of carriers between neighboring QD’s plays
no significant role in the PL temperature dependence.
ACKNOWLEDGMENTS
C.L. thanks C. Jagadish for helpful discussions. Part of
this work was sponsored by the Australian Research Council
and by the Jet Propulsion Laboratory, under a contract with
the National Aeronautics and Space Administration.
1
G. Wang, S. Fafard, D. Leonard, J. E. Bowers, J. L. Merz, and P.
M. Petroff, Appl. Phys. Lett. 64, 2815 共1994兲.
2
S. Raymond, S. Fafard, P. J. Poole, A. Wojs, P. Hawrylak, S.
Charbonneau, D. Leonard, R. Leon, P. M. Petroff, and J. L.
Merz, Phys. Rev. B 54,11548共1996兲.
3
W. Yang, R. Lowe-Webb, H. Lee, and P. C. Sercel, Phys. Rev. B
56,13314共1997兲.
4
M. Braske
´
n, M. Lindberg, M. Sopanen, H. Lipsanen, and J.
Tulkki, Phys. Rev. B 58, R15 993 共1998兲.
5
Y. Zhang, M. D. Sturge, K. Kash, B. P. van der Gaag, A. S.
Gozdz, L. T. Florez, and J. P. Harbison, Superlattices Micro-
struct. 17, 201 共1995兲; Phys. Rev. B 51,13303共1995兲.
6
R. Leon, S. Marcinkevic
˘
ius, X. Z. Liao, J. Zou, D. J. H. Cock-
ayne, and S. Fafard, Phys. Rev. B 60, R8517 共1999兲.
7
R. Leon, C. Lobo, J. Zou, T. Romeo, and D. J. H. Cockayne,
Phys. Rev. Lett. 81, 2486 共1998兲.
8
H. Benisty, C. M. Sotomayor-Torre
`
s, and C. Weisbuch, Phys.
Rev. B 44,10945共1991兲.
9
U. Bockelmann, Phys. Rev. B 48,17637共1993兲.
10
R. Leon, S. Fafard, P. G. Piva, S. Ruvimov, and Z. Liliental-
Weber, Phys. Rev. B 58, R4262 共1998兲.
11
J. D. Lambkin, D. J. Dunstan, K. P. Homewood, L. K. Howard,
and M. T. Emeny, Appl. Phys. Lett. 57, 1986 共1990兲.
12
G. Bacher, H. Schweizer, J. Kovac, A. Forchel, H. Nickel, W.
Schlapp, and R. Lo
¨
sch, Phys. Rev. B 43, R9312 共1991兲.
13
S. Fafard, D. Leonard, J. L. Merz, and P. M. Petroff, Appl. Phys.
Lett. 65, 1388 共1994兲.
14
S. Paul, J. B. Roy, and P. K. Basu, J. Appl. Phys. 69, 827 共1991兲.
15
S. Fafard, S. Raymond, G. Wang, R. Leon, D. Leonard, S. Char-
bonneau, J. L. Merz, P. M. Petroff, and J. E. Bowers, Surf. Sci.
362, 778 共1996兲.
16
Z. Y. Xu, Z. D. Lu, Z. L. Yuan, X. P. Yang, B. Z. Zheng, J. Z.
Xu, W. K. Ge, Y. Wang, J. Wang, and L. L. Chang, Superlat-
tices Microstruct. 23, 381 共1998兲.
17
A. Patane
`
, M. Grassi Alessi, F. Intonti, A. Polimeni, M. Capizzi,
F. Martelli, M. Geddo, A. Bosacchi, and S. Franchi, Phys. Status
Solidi A 164, 493 共1997兲.
18
X. Z. Liao, J. Zou, D. J. H. Cockayne, R. Leon, and C. Lobo,
Phys. Rev. Lett. 82, 5148 共1999兲.
19
M. Vening, D. J. Dunstan, and K. P. Homewood, Phys. Rev. B
48, 2412 共1993兲.
20
The hole mobility in In
x
Ga
1⫺ x
As peaks between 70 and 100 K,
depending on the relative contributions of ionized impurity scat-
tering, polar optical phonon scattering, acoustic phonon scatter-
ing, and alloy scattering 关see S. Adachi, Physical Properties of
III-V Semiconductor Compounds 共Wiley, New York, 1992兲兴.
21
S. Marcinkevic
˘
ius and R. Leon, Phys. Rev. B 59, 4630 共1999兲.
22
B. K. Ridley, Phys. Rev. B 50, 1717 共1994兲.
23
G. C. Crow and R. A. Abram, Semicond. Sci. Technol. 14,1
共1999兲.
24
J. Tersoff, Phys. Rev. Lett. 81, 3183 共1998兲.
PRB 60
16 651INHIBITED CARRIER TRANSFER IN ENSEMBLES OF...
