вход по аккаунту



код для вставкиСкачать
Intersublevel electroluminescence from
In0.4Ga0.6As=GaAs quantum dots in
quantum cascade heterostructure with
GaAsN=GaAs superlattice
C.H. Fischer, P. Bhattacharya and P.-C. Yu
Infrared light emission is demonstrated from In0.4Ga0.6As selforganised quantum dots in a quantum cascade structure incorporating
a GaAs=GaAsN chirped superlattice. A TE polarised emission peak
is observed at 22 mm, corresponding to the calculated first-excited
to ground-state transition. Other emission peaks also correspond to
predicted values.
Introduction: While quantum cascade devices [1, 2] have hitherto
been designed with quantum well active regions, it is of interest to
investigate quantum dot active regions [3, 4] for the following
reasons. The polarisation selection rules allow surface-normal (vertical) emitting devices. Large tunability of the emission wavelength can
be achieved by tailoring the growth conditions during island formation and the quantum dot heterostructure. Discrete energy levels
reduce phonon coupling [5] in self-organised quantum dots compared
to quantum wells, thus promising high-temperature operation of
intersublevel emitters. The calculated and measured lifetimes of
electrons in the excited states are of the order of 100 ps [6]. It has
been established that injected electrons preferentially occupy the
excited dot states and wetting layer states at high temperatures due
to the higher density of states in these regions [7, 8], suggesting that
population inversion can be readily obtained. In the present study, we
have designed and characterised strain-compensated unipolar quantum-dot-cascade structures, using the lowest energy levels in the dot.
the ground state in one dot layer to the first-excited state of the next,
thus producing 57 meV photons as the electron relaxes to the ground
state in each successive dot layer. Dilute nitride GaAsN was selected for
the quantum well in the CSL because it provides both large conduction
band offset and tensile strain compensation of the quantum dot layers.
The final structure is GaAs0.99N0.01(67, 54, 37, 23Å)=GaAs(35, 22, 39,
60, 63Å) well=barrier thicknesses. A self-consistent solution to the
Schrödinger–Poisson equations for this structure at 60 mV bias is
shown in the inset of Fig. 1 with eigenstates from the 8-band k p
quantum dot simulation superimposed on the left. First-order perturbation theory was used to verify that the introduction of a chirped
superlattice does not significantly alter the eigenstates in the dot. A
single cascade period consists of the quantum dot active region
followed by the n-type (2 1017 cm3) CSL injector, both of which
are described above. The cascade structure is repeated 11 times on an
nþ (1019 cm3) substrate with an nþ top contact.
Experimental: The device heterostructures were grown by molecular
beam epitaxy with a Unibulb plasma nitrogen source. The quantum
dots were grown at 520 C, and the CSL regions were grown at 450 C.
Following growth, the sample is annealed in situ at 650 C for 20 min
under As4 overpressure. Surface atomic force microscopy (AFM) was
performed on similarly grown samples in which the dot layer was
grown on one or more periods of the cascade heterostructure. No
significant variation between successive periods is observed, indicating that the tensile strain in the CSL does not significantly alter the
size or shape of the dots. Edge emitting devices were fabricated with
the heterostructures using standard optical lithography, n-ohmic, and
wet etching techniques. Ridge widths vary from 50 to 300 mm, and the
length varies from 0.5 to 2 mm. Vertical emitting 300 mm diameter
mesas with a ring top-contact were also fabricated by similar techniques, and will be discussed briefly in the concluding remarks.
Device design: The design of the quantum dot devices is nearly
identical to conventional quantum cascade lasers [1–3], but with a
strain-coupled In0.4Ga0.6As=GaAs quantum dot active region. Dot
layer thicknesses of 7 and 5 ML with a 15 Å GaAs barrier are chosen
to produce dots the characteristics of which are repeatable and well
known. Atomic force microscopy on such dots shows the dot base
width and height are 15 and 8 nm, respectively, and the density is
3 1010 per cm2. An 8-band k p simulation [9, 10] based on these
values predicts intersublevel transitions of 57 and 87 meV, respectively, for the first- and second-excited state to ground transitions.
These energy transitions agree with previously reported experimental
results [11] and are indicated along with measured electroluminescence data.
energy, meV
Fig. 2 Polarisation dependent spectral measurements for broad-area
device at 18K showing dominant TE polarisation
intensity, a.u.
Theoretically predicted energies for first- and second-excited to ground state
transitions are indicated (as E21 and E31) for comparison. The broad, high energy
(115 meV) peak possibly results from wetting-layer to ground-state transitions
position, nm
T = 90K
pulsed bias
current, mA
Fig. 1 Light–current characteristics from edge-emitting broad area
(300 mm 2 mm) device at 90K
A distinct injector turn-on is evident, illustrated by dashed lines
Inset: Solution to coupled Schrödinger–Poisson equations for chirped superlattice
at 60 mV=period with quantum dot states from 8-band k p model superimposed
on left
A chirped superlattice (CSL) injector region was designed to both
compensate strain from the quantum dots and transmit electrons from
Results and discussion: Fig. 1 shows light–current (L–I ) measurements made on edge emitting devices at 90K with a closed-cycle
helium cryostat, pulse generator, lock-in amplifier, and liquid-nitrogencooled HgCdTe detector. Since the injector region is designed to
operate under a bias of 60 mV per period (660 mV over the entire
device structure), the ‘turn-on’ slightly above this bias, as illustrated
by the dashed lines, is expected. The polarisation-resolved electroluminescent spectra for the same device at 18K, as measured by FTIR,
are shown in Fig. 2 along with the predicted quantum dot energy
transitions described above. Data from 500 sequential FTIR scans
were averaged to reduce noise, and ten such sets of data were
collected for each polarisation. Background spectra were collected
with the device unbiased and subtracted from the biased data. All TE
polarised outputs show the same three emission peaks, while none of
the TM polarised outputs exhibit these features. Some data sets show
slight (less than 1%) variations in magnitude across the entire spectral
range, likely due to changes in the temperature of the optics. These
ELECTRONICS LETTERS 16th October 2003 Vol. 39 No. 21
variations are clearly distinguishable from the actual signals as the
resulting spectrum shows a trend similar to the background data. Data
shown are averages of all data sets for each polarisation. Two peaks
(57 and 90 meV) agree well with first- and second-excited to groundstate transitions, while a third peak (115 meV) arises, most likely,
from wetting-layer to ground-state transitions. The dominant TE
polarisation agrees with theory [6]. Unlike in quantum wells, the
oscillator strength in quantum dots arises from the central cell part of
the wavefunction. Strong biaxial strain in dots leads to intermixing of
bands for TE polarised light, which results in a larger momentum
transfer matrix element for TE polarised light. This property of
quantum dots has been observed previously in bipolar quantum dot
intersublevel devices [12]. Total power from the 57 meV peak is
estimated to be 0.7 nW. Total integrated output power for all three
peaks is estimated to be 90 nW.
Spectra similar to the TE polarised signal in Fig. 2 have also
been observed from surface emitting devices at 18K, but the signal
was much weaker due to the smaller device area. These are, therefore,
not shown here.
The presence of electroluminescence peaks beyond the designed E21
intersublevel transition (’ 57 meV), with peak and integrated intensity
larger than that of E21, suggests that a significant number of injected
carriers are not captured into the first dot layers. Instead, they are
transported through the chirped superlattice and occupy the higher
energy states (barrier and wetting layers) of subsequent dot layers.
These electrons subsequently relax to the lower energy quantum dot
levels and give rise to the high energy transitions.
Acknowledgment: The work is supported by the Air Force Office of
Scientific Research (DoD MURI program) under Grant F4962000-1-0328.
# IEE 2003
Electronics Letters Online No: 20030970
DOI: 10.1049/el:20030970
6 August 2003
C.H. Fischer, P. Bhattacharya and P.-C. Yu (Solid State Electronics
Laboratory, Department of Electrical Engineering and Computer
Science, University of Michigan, Ann Arbor, MI 48109-2122, USA)
E-mail: [email protected]
FAIST, J., et al.: ‘Quantum cascade laser’, Science, 1994, 264, (5158),
pp. 553–556
2 See, for example, papers in IEEE J. Quantum Electron., 2002, 38, (6),
pp. 510–693
3 HSU, C.F., et al.: ‘Intersubband quantum-box semiconductor lasers’, IEEE
J. Sel. Top. Quantum Electron., 2000, 6, (3), pp. 491–503
4 APALKOV, V.M., and CHAKRABORTY, T.: ‘Optical properties of a quantumdot cascade structure’, Physica E, 2002, 14, (3), pp. 294–298
5 URAYAMA, J., et al.: ‘Observation of phonon bottleneck in quantum dot
electronic relaxation’, Phys. Rev. Lett., 2001, 86, (21), pp. 4930–4933
6 KOCHMAN, B., et al.: ‘Absorption, carrier lifetime, and gain in InAs–GaAs
quantum-dot infrared photodetectors’, IEEE J. Quantum Electron., 2003,
39, (3), pp. 459–467
7 URAYAMA, J., et al.: ‘Temperature-dependent carrier dynamics in selfassembled InGaAs quantum dots’, Appl. Phys. Lett., 2002, 80, (12),
pp. 2162–21641
8 MATTHEWS, D.R., et al.: ‘Experimental investigation of the effect of
wetting-layer states on the gain-current characteristic of quantum-dot
lasers’, Appl. Phys. Lett., 2002, 81, (26), pp. 4904–4906
9 JIANG, H.T., and SINGH, J.: ‘Strain tensor and electron and hole spectra in
self-assembled InGaAs=GaAs and SiGe=Si quantum dots’, Physica E,
1998, 2, (1–4), pp. 614–618
10 JIANG, H.T., and SINGH, J.: ‘Strain distribution and electronic spectra of
InAs=GaAs self-assembled dots: an eight-band study’, Phys. Rev. B,
1997, 56, (8), pp. 4696–4701
11 KLOTZKIN, D., et al.: ‘Electron intersubband energy level spacing in selforganized In0.4Ga0.6As=GaAs quantum dot lasers from temperaturedependent modulation measurements’, J. Vac. Sci. Tech. B, 1999, 17, (3),
pp. 1276–1280
12 KRISHNA, S., et al.: ‘Room-temperature far-infrared emission from a selforganized InGaAs=GaAs quantum-dot laser’, Appl. Phys. Lett., 2000, 76,
(23), pp. 3355–3557
ELECTRONICS LETTERS 16th October 2003 Vol. 39 No. 21
Без категории
Размер файла
198 Кб
Пожаловаться на содержимое документа