Download ppt

Tuning the properties of
exciton complexes in selfassembled GaSb/GaAs
quantum rings
M. Ahmad Kamarudin, M. Hayne,
R. J. Young, Q. D. Zhuang,
T. Ben, and S. I. Molina
Type-II self-assembled GaSb/GaAs nanostructures have
been grown by molecular-beam epitaxy and studied by
atomic-force microscopy, transmission electron
microscopy, and power-dependent
magnetophotoluminescence.
Nanostructures on the sample surface are found to be
entirely dotlike, while capped nanostructures are
predominantly ringlike. Moreover, an in situ anneal
process applied after thinly capping the dots is shown to
enhance the severity of the rings and relax the strain in
the matrix in the proximity of the GaSb, resulting in a
change to the spatial configuration of the exciton
complexes and their optical properties.
Outline
•
•
•
•
Introduction
Sample
Morphology and features
Experiment
1. zero field photoluminescence
2. magnetophotoluminescence
• Discussion
• Conclusions
Introduction
• GaSb/(Al)GaAs can to extend the
absorption spectrum of GaAs solar cells
beyond 1 mm.
• Provide deep confining potentials capable
of room-temperature charge storage for
memory applications
Motive
• Compare the differences of GaSb
quantum ring by MOCVD and MBE.
1. Morphology
2. Low temperature PL
GaSb/GaAs quantum dot
(a) PL peak transition energy as a function of magnetic field at different laser powers
measured at 4.2 K. (b) A depiction of the corresponding electron trajectories
represented by solid lines for bound states and dotted lines for free particles as
the sample goes through a density-dependent Mott transition. (c) The values of the
exciton radius squares and the exciton binding energy circles as a function of
excitation power, obtained from the fits in (a).
PHYSICAL REVIEW B 77, 241304
InP/GaAs quantum dots
e 2 aB2 B 2
E  E0 
for B  Bc (1)
8m
2
eB
E  E0 

2 maB2 2 m
for B  Bc (2)
Bc 
2
eaB2
The measured peak shift as a function of magnetic field at the excitation power
of about 10−4 W cm−2. The crossover from Coulombic confinement
to magnetic confinement occurs at 5.9 T and corresponds to an
effective Bohr radius of 15 nm. The extrapolation of the linear slope
measured at high field to zero field gives the binding energy.
PHYSICAL REVIEW B 80, 205317
Sample Growth
2.1 monolayers 490 ◦C
0.3 ML s-1 GaSb Sb
100 nm 500 ◦C
1 MLs-1 GaAs
~9 nm cold cap GaAs
430 ◦C.
2.1 monolayers 490 ◦C
0.3 ML s-1 GaSb
GaAs
Sample B was subjected to a 2-min
growth interruption under As2 flux at
580 ◦C before being capped with
100 nm of GaAs
Microscopy
A
B
Atomic-force-microscopy images of the surfaces of the two samples
Sizes and area densities of the GaSb QD’s on the surfaces of the two samples
TEM images in dark-field 002 imaging conditions. In sample A, the dark region above
the GaSb nanostructures indicates the presence of strain.
High resolution TEM images of a buried single-lobe nanostructure in
sample A (a), and a double-lobe nanostructure in sample B (b).
The dissolution effect of
capping the nanostructures
LOW-TEMPERATURE
MAGNETOPHOTOLUMINESCENCE
• Low-temperature (≦ 4.2 K)
• Frequency-doubled diode-pumped solid-state
laser (532 nm)
• 30-cm focal length spectrometer was
combined with a Peltier-cooled InGaAs diode
array
• Spot diameter on the sample was ∼2 mm
Low-temperature photoluminescence (PL) data at zero magnetic field
Sample A (no anneal) is in blue and sample B (anneal) is in red.
Spectra: the quantum ring (QR), wetting layer,and GaAs peaks are labeled.
Simplified band diagram showing spatially indirect excitons
with electrons in the GaAs matrix and holes in the GaSb
valence band of the WL and QD. The z axis denotes the
growth direction and the QDs are distributed in the xy plane.
PHYSICAL REVIEW B 77, 241304
Band energy diagram deduced from the PL transition energies obtained at 15 mW
and 20 K taking into account electron accumulation.
APPLIED PHYSICS LETTERS 91, 263103
Explain the difference in emission
energy
1. a reduction in QR charging in sample B
due to lower unintentional p-doping
compared with sample A
2. an increase in the size of the
nanostructures in sample B when
compared to sample A
3. a change in the composition induced by
the different capping conditions for the two
samples.
At low power photoexcited
holes are captured by
unoccupied carbon
acceptors, C−, while
photoexcited electrons
recombine with the holes in
the dots, reducing the hole
population and causing a red
shift in the PL. At high laser
powers all the carbon
acceptors are occupied, so
photoexcited holes are
rapidly captured by the QD,
causing a blue shift in the
PL.
PHYSICAL REVIEW B 70, 081302
Low-temperature photoluminescence (PL) data at zeromagnetic field
sample A (no anneal) is in blue and sample B (anneal) is in red.
For this reason, the use of the term “exciton” in
this paper should not necessarily be taken to imply
single electron-hole pairs, but should be more
broadly interpreted to include charged excitons and
other exciton complexes.
The inset schematically shows the size and displacement of the electron
from the GaSb (shaded region) for (a) WL, (b) QD under low intensity illumination
and (c) QD under high intensity illumination. The number of holes in (b) and (c)
are intended to signify a greater or lesser occupation, rather than actual
occupancies.
Physica B 346, 421
2. an increase in the size of the nanostructures in
sample B when compared to sample A
• Confinement energies for holes are much less sensitive
to size effects than for electrons, and the TEM data
indicate no systematic change in QR base length or
height between samples A and B
• Because sample B was annealed, the strain in the GaAs
close to the quantum rings were smaller than the sample
A.
Magnetophotoluminescence
Hamiltonian for a single electron in a uniform magnetic field B

P  eA(r) 
H
2
2me*
gμB
 eVe (r) 
SB

e
mB 
2me
P
e A
e
gm B
H
 eVe  * P  A  A  P  

SB
*
*
2me
2me
2me

2
2
2
Time independent perturbation
2
P
E0 
 eVe
*
2me
guB
1
S  B   guB B

2
e
P  A  A  P 
HI 
*
2me
(P  A )   i( A )  i(  A )  A   
  iA     A  P 
e
e
2A  P   * L  B
HI 
*
2me
2me
1
A  B ( yi  xj)
2
e
e
e
HI 
LB 
BLZ 
Bml
*
*
*
2me
2me
2me
E 
2
0

j ( 1)
1HI j
E E
0
0
2
0
j
0
2
H II
2
e A

*
2me
2
1
A  B ( yi  xj)
2
2
2
2
e 1 2 2
e B
2
2
2
E 
B (y  x ) 
(y  x )
*
*
2me 4
8me
1
0
2
2
B
ea B
E  E0 
8m
2

2
eB
E  E0 

2
2maB 2m
 me4
En  2 2
2 n
2


2maB2
electron and hole pair
energy level
2
aB 
me2
precession
e
H 
LB
*
2me
e
e

BLZ 
Bml
*
*
2me
2me
e 2 aB2 B 2
E  E0 
for B  Bc (1)
8m
2
eB
E  E0 

2 maB2 2 m
for B  Bc (2)
Bc 
2
eaB2
E0 is the zero-field PL energy
aB is the exciton Bohr radius
μ is the reduced exciton mass
Quantum-ring PL peak position as a function of magnetic field.
The lines are fits to Eqs. (1) and (2). It can be seen that the field-induced
shift of the PL is much larger for sample B than sample A, and that
for sample A the field dependence is highly parabolic at low laser
power (b) and more linear at high laser power (a). For sample B, the
opposite trend is observed (d).
Variation of the magnetic-field, B,
induced shift of the QR PL vs laser
excitation power, P. Note that an
increasing B-shift with increasing
laser power indicates increasing
excitonic binding with increasing
laser power
The inset shows the exciton Bohr
radius vs P. For sample A
(no anneal) and P < 10 mW,
the Bohr radius is too small to
be measured with the magnetic
fields available.
Equations (1) and (2) may be used to
determine the useful parameters
aB and μ by fitting the PL peak energy
data over the whole field range in a
single operation.
Sample A μ = 0.30 ± 0.02m0
Sample B μ = 0.096 ± 0.005m0
(m0 is the free-electron mass)
e 2 aB2 B 2
E  E0 
for B  Bc (1)
8m
2
eB
E  E0 

2 maB2 2 m
for B  Bc (2)
Bc 
2
eaB2
Discussion
• Sample A (no anneal) has a smaller opening in the center of
the QR than sample B, with a higher Sb content.
• Since the electron is weakly bound to the hole in the
QR, when laser power is increased, electron-electron
interactions (screening) dominate over electron-hole
interactions, resulting in a smearing out of the
electron wave function, which increases the B-field
shift and Bohr radius
• Sample B (anneal) has a larger opening in the QR
with a lower Sb content
• the close proximity of the electron to the QR means
that the electron-hole interaction dominates, and as
the photoexcited holes charge the QR with increasing
laser power, the excitonic binding is increased, not
reduced. This decreases the size of the B-field shift
and the Bohr radius.
Conclusions
Annealing of the second sample resulted in changes
in the ring morphology that effect the spatial
configuration of the excitons, leading to quite different
optical properties in the two samples.
We hope that this will stimulate further
theoretical and experimental work, leading to the
observation of AB oscillations in GaSb QR’s in the
future.