Transcript NAK_P25_NGS2007

Nano and Giga Challenges
in Electronics and
Photonics
P-25
Quantum-Dot –Array Based Terahertz Detectors
Tempe, Arizona, March 12-16, 2007
N. A. Kabir1, J. Song1, A. Markelz2, T. Morimoto3, Y. Ujiie3
3
3
3
3
1
N. Yumoto , K. Sudou , N. Aoki , Y. Ochiai , & J. P. Bird
1: Department of Electrical Engineering, University at Buffalo, the State University of New York, Buffalo, NY 14260, USA
2: Department of Physics, University at Buffalo, the State University of New York, Buffalo, NY 14260, USA
3: Department of Electronics and Mechanical Engineering, Chiba University,1-33 Yayoi-cho, Inage-ku, Chiba 263 8522,
Japan
We examine the terahertz conductivity response of lithographically
defined quantum dot arrays as a function of temperature, dot size and
photo excitation. Quantum dot structures have been explored as
possible compact sources and detectors of THz radiation. Here we
consider a lithographically defined array with high uniformity and high
duty cycle to increase FIR optical density. Specifically we examine the
cross over from continuum to discrete response as the quantized
energy level spacing becomes commensurate with the FIR.
Interaction of THz photons with a 2DEG in a semiconductor provides the
basis for a number of different THz-detection schemes.
One of the simplest approaches makes use of free carrier absorption to heat
the electron gas relative to the lattice to change the conductivity. But it
requires low-temperature (4.2 K) operation and does not provide frequency
sensitivity.
One way to realize such sensitivity is to make use of an appropriate
electronic transition between the discretely-quantized states of quantum dot
or quantum well, however, this again imposes the restriction of lowtemperature operation.
Samples are lithographically defined using GaAs/Al0.3Ga0.7As and
InAs/AlSb high-mobility, modulation doped two dimensional electron
gas (2DEG )wafers. Quantum-dot arrays are fabricated using e-beam
lithographically and wet chemical etching. Measurements of the
complex conductivity response at THz frequencies are made using
terahertz time domain spectroscopy (THz TDS).
An alternative approach that should overcome these issues, allowing for
higher temperature and frequency-specific operation, uses THz radiation to
excite collective plasma oscillations in a high-mobility 2DEG, most
commonly achieved by forming a metal grating on the heterostructure
surface.
Abstract
1.
Background
2.
1) By exploiting the advantages of epitaxial growth, distance to the 2DEG
can be reduced to a value of ~10 nm, about an order of magnitude
smaller than the grating-2DEG separation in conventional devices,
without any significant degradation of the 2DEG mobility. Given the
exponential decay of the scattered electromagnetic-wave amplitude
with distance from the grating, this is potentially a large effect
 contactless
measurement of
frequency dependent
conductivity
2) The QD-array is characterized by its own plasmon excitations (grating
plasmons) and the frequency of these should also lie in the THz range,
similar to the plasmon mode in the adjacent 2DEG. When the frequency
of the incident radiation is close to that of the grating plasmon,
resonant excitation of these plasmons should cause the amplitude of
the scattered electromagnetic field in the grating layer to increase
dramatically and even exceed that of the external field
 measuring the
complex conductivity
response and assuming
a Drude model one can
directly access the
momentum relaxation
rate
Phase vs Frequency
1.1
Transmission
0.9
5K
10K
25K
50K
75K
100K
150K
200K
250K
300K
0.8
0.8
1.2
Frequency (THz)
1.6
2
0.4
0.8
1.2
Frequency (THz)
1.6
o  8.7  10 rad / s
o
f0 
 0.14THz
2
0.8
1.2
Frequency (THz)
1.6
AlGaAs: 30 nm
(GaAs BIG QD sample shown here)
GaAs : 800 nm
GaAs 2DEG sample –
m* = 0.067 m0
ns = 2.5 X 1015 m-2 (RT)
2DEG
GaAs Substrate : 600 μm
Array of QD Samples
6.
Fermi-energy calculation
Assuming parabolic potential
1 * 2 W2
EF  m o
2
4
8 EF
o 
m*W 2
En(ky)
2
-0.35
EF
-0.4
0.4
0.8
1.2
Frequency (THz)
k F2
37
EF 

5.73
X
10
XnS
*
2m
Here,
nS  2.5 X 1015 m 2
Absorbance
2
1.6
2
Results – BIG (500nmX500nm) QD
8.
ky
 EF  9meV
Calculations
9.
1.4
n e
n e
q
2
p  S * .
 S* .
m  (qt )
m
L r  0
• No illumination effects observed
1.2
Frequency (THz)
7.88  109
p 
2 L
for L=500nm
1
GaAs 2DEG
0.9
0.8
0.8
0.6
Fmin1_GaAs_BIG
Fmin2_GaAs_BIG
Fmin_InAs
Fmin_GaAs_SML
 p  7.88 10 rad / s
12
0.4
p
fp 
 1.25THz
2
1
10
100
Temperature (K)
• At low T: true “dot” response
for L=200nm
 p  12.44 10 rad / s
12
p
fp 
 1.98THz
2
Calculations
InAs 2DEG
1000
• True “dot” response is seen at low T and washes out at above ~100K due
to electron-phonon scattering effect
0.7
• At High T (above ~100K): response
determined by electron-phonon
scattering – as Ns changes
drastically for both samples
11.
• Inflection points from the phase data shows absorbance due to plasmon
oscillation in the dots – visible only in GaAs BIG (~500nmX500nm) QD
sample
2nd harmonic
8.6
for W=200nm
o  2.17  1012 rad / s

f 0  o  0.35THz
2
n-AlGaAs: 40 nm
2
for W=500nm
11
GaAs BIG QD = 500nmX500nm
GaAs SML QD = 200nmX200nm
InAs QD
= 200nmX200nm
Occurence of minima
2
o 
5K
10K
25K
50K
75K
100K
150K
200K
250K
300K
0.4
2
Oscillator and plasmon frequency calculation
8 EF
m*W 2
0.9
0.7
Results – SMALL (200nmX200nm) QD
7.
1
0.8
-0.4
0.4
2nd harmonic
-0.35
0.7
Drawbacks of existing detectors
3.
5K
10K
25K
50K
75K
100K
150K
200K
250K
300K
-0.3
Phase
Transmission
1
5) Distance to 2DEG can not be reduced below a few hundred nm, as it deteriorates
the 2D electron mobility and the frequency resolution of the plasmon system
restricting the coupling efficiency of the metal grating and detection sensitivity
-0.25
1.2
5K
10K
25K
50K
75K
100K
150K
200K
250K
300K
-0.3
4) Fourier harmonics of the electromagnetic field are attenuated by the time
they reach the 2DEG
Phase vs Frequency
Transmission vs Frequency
-0.25
1.1
3) Amplitude of the induced electric field penetrating to 2DEG cannot exceed the
externally-incident radiation due to the total electric field inside the metal grating
fingers being equal to zero at all points due to screening
Phase
1.2
15-DOT CRYSTAL
GaAs CAP: 5 nm
Schematic of THz TDS
5.
Ns-Hall (Xe16 /m2)
Transmission vs Frequency
2) Electromagnetic wave that is scattered by the grating gate corresponds to an
evanescent mode whose amplitude decays exponentially with increasing distance
from the gate with degrading sensitivity
 coherent detection
Motivation
4.
1) Metal gate induces a periodic modulation of the normally-incident external THz
radiation, which lies in the submillimeter range and is much larger than the grating
period
 ultra-fast optical
technique that allows us
to measure the dielectric
response in the range 5 –
85 cm-1
In order to overcome the problems noted above, we propose to
investigate the characteristics of a new class of plasmonic THz detectors
in which the metal grating gate is replaced with nanostructured arrays of
quantum dots with the following advantages:
10.
The critical issue in plasmon-based detection schemes concerns an efficient
9-DOT CRYSTAL
coupling of the external electromagnetic radiation and the 2D plasmons
which suffers due to the following reasons:
8.4
0.6
8.2
0.5
• No proof of carrier transitions between different energy sub-bands but
there is definite presence of 2nd order harmonic oscillations
8
7.8
0.4
7.6
1
10
100
• GaAs SMALL QD response seems to indicate that it is always depleted
0.3
0.2
1
10
Temperature (K)
100
Plasmonic detection
• InAs QD shows a weaker confinement with a constant Ns
12.
Conclusion
NANOELECTRONIC MATERIALS & DEVICES RESEARCH GROUP (NoMaD)
Department of Electrical Engineering, University at Buffalo