3rd Korea-USA NanoForum, Tower Hotel Seoul, April 3-4, 2006
Nano-optoelecrtronics based on III-V
Semiconductor Quantum Dots
Jungil Lee
Nano Device Research Center, KIST
[email protected]
KIST
- Founded 1966, gouv. affilliated
- 4 Res. Div.s & 1 Res. HQ(future technol.)
Materials, Systems, Environ., Bio-sci
- 654 (422(306 Ph.D.)researchers)
- 130MU$/yr, ‘04
- Future Technol. HQ
Micro Systems Res. Ct.
Nano Device Res. Ct.
Simulation (supercom., cluster type)team
Nano Device Res. Ct.
-
Mag. Mater. group + Semicon. Group
(Spintronics)
Semicon. group.
Joo Sang Lee, Young Ju Park (MIT lab)
Jin Dong Song (MBE)
Won Jun Choi (QD LD, QDIP)
Ilki Han (Hi pwr LD, SLD, QCL) (NRL)
Woon Jo Cho (nc Si)
Jung Il Lee (electrical charac.)
~ 20 studs & 1 PD
Collaborators- domestic
SRC:
- QSRC (TWKang, Dongguk U.), GaN
- QPSI (YPLee, Hanyang U.), polymer PC
NRL:
- VCSEL (YHLee, KAIST)
- QD Technol (SKNoh, KRISS)
- Ultrafast Phenom (DSKim, SNU)
- Compound Semi Epi (EJYoon, SNU)
- QD Opt Dev (DHLee, CNU)1.5QD, SOA
- QD PL (YHCho, Chungbuk NU)
Collaborators- international
Germany: Wuerzburg
UK: Sheffield
France: CNRS(IEF, LPN, LSP, LEOM, IMEP,
LAAS)
China: CAS-SITP, -IoSemi
Japan: U.Tokyo, NICT
USA:
Programs
Dual: QCL –’07
Nano R&D (QM D/Wr Technol, 7U+KIST+3Co):
- QD(LD, SOA, Mod)
- PC(passive, active) -’11
ETRI: 1.55 um QD LD
ABITD: QSIP (KIST, KRISS, KAIST) -’04
Frontier(TND): Opt Interconnect (YTLee,
GIST)
Growth Engin: GaN LED (KOPTI)
Contents
I. QD vs. QW
II. QD growth (ALE)
III. QD LD
IV. QD SLD
V. QD IP
VI. Summary
VII.Prospects
I. QD vs. QW
‰ Carrier localization & Discrete DOS:
- Low leakage currents of LDs
- Small active volume ; low transparent current density of LDs
- Large activation (thermalization) energy ;
High temperature operation of LDs and IR-PDs
Stable operation of high power laser at high temp.
- Low refractive index change by injection current ;
Small α-parameter -> low chirping (high speed operation)
small filamentation (high power LD)
‰ Inhomogeneous broadening due to size fluctuation:
- Large gain bandwidth ; SOA, ECL, SLD
II. QD Growth (ALMBE)
ALMBE mode (ALE mode)
ALE:Atomic Layer epitaxy
n
GaAs
GaAs
As
0.6
0.4
0.2
In
DWELL 구조
7.5 nm In0.2Ga0.8As
1.2 nm In0.2Ga0.8As
0.0
900
InAs QD
λ= 1250 nm
D: 1.02 x 1011 /cm2
In0.5Ga0.5As QD
λ= 1150 nm
64 meV
T = 300 K
1400
650 mW
1200
400 mW
InAs/GaAs ALE QD
14860A
RT
1000
34 meV
200 mW
100 mW
50 mW
1000 1100 1200 1300 1400
Wavelength (nm)
PL intensity (A.U.)
Dots
PL Intensity (a.u.)
• Wetting layer is ignorable
• Relatively large dot size
(> ~30 x ~10 nm)
• Wavelength ( more than 1500 nm is
possible)
800
600
94.3meV
400
200
0
-200
12000
13000
14000
15000
wavelength(A)
Bandgap engineering of QD : 1150 ~ 1500 nm
Detection wavelength tuning possible
16000
QD Growth
SK mode dot
ALE mode dots
• Wetting layer is inevitable
• Wetting layer is not necessary
• Relatively small dot size
• Relatively large dot size
(≤ ~30 x ~10 nm)
(≤ ~60 x ~10 nm)
• Wavelength (1000 ~ 1300 nm + α) • Wavelength ( more than 1500 nm is
possible)
Dots
GaAs
Weting layer
Dots
N periods
of layers
GaAs
GaAs
In
As
AFM images
Sample #A(ALE)
Sample #B(S-K)
Dot density, width and height of the sample #A are ~4.1x1010/cm2, ~40.8 nm, and ~7.2 nm, respectively,
In the case of the sample #B, 13.4x1010/cm2, ~31.0 nm, ~2.6 nm, respectively.
HR-TEM images (ALE dot)
~32o
{136} facet formation
3
nm
3
nm
We can observe indium diffusion
into GaAs below the QDs
Indium can diffuse into GaAs matrix below the QDs during growth interruptions, and
this helps the QDs
HR-TEM images(SK dot)
~ 23 o
The QDs have flat bottom-shape
Wetting layer thickness
Sample #A(ALE)
Thickness of WL : ~ 2.1 nm
Sample #B(S-K)
Thickness of WL : ~ 4 nm
Reduced thickness of wetting layers in the QDs grown by ALMBE was
predicted by Guryanov et al. Surf. Sci. 352-354, 651 (1996).
Normalized PL intensity (arb. units)
PL spectra
1.2
1.0
T = 300 K
Sample #B
Sample #A
Sample #B has only 1 level.
0.8
0.6
0.4
~64 meV
~29 meV
(x ~20)
0.2
0.0
0.9
Sample #A has a 2 peaks.
(Ground and 1st excited level)
1.0
1.1
1.2
Wavelength (µm)
1.3
FWHM of sample #A (~29 meV)
is ~ half of that of sample #B
(~64 meV).
This implies that size distribution
of sample #A is more uniform
Than sample #B.
Temp. depen’t PL spectra
1.06
1.04
1.02
1.00
0
50
100 150 200 250 300
Temperature (K)
74
72
70
68
66
64
62
60
58
1.20
1.18
1.16
1.14
1.12
1.10
0
PL peak FWHM (meV)
Ground level
Sample #B(S-K)
1.22
PL peak energy (eV)
1.08
40
38
36
34
32
30
28
26
24
22
(b)
PL peak FWHM (meV)
PL peak energy (eV)
(a) Sample #A(ALE)
50 100 150 200 250 300
Temperature (K)
Temperature dependence of PL peak energy and PL peak linewidth of (a) the sample #A,
and (b) the sample #B. In the case of the sample #A, only ground level is considered.
It is noteworthy that the PL linewidth of the sample #A is insensitive to cryostat temperature
due to lack of wetting effect, that is, reduced thickness of wetting layers and/or uniform formation
of QDs in ALMBE
APL 88, 133104(2006)
Hi quality QD
S-K dot vs ALE dot
7
e
1000
ALE(G.S) : 2.2ns
900
15 K
800
GaAs 500 A
3 stack InAs ALE dot
6
e
SK(G.S): 800ps
5
e
4
e
3
e
PL Intensity (arb. uints)
Normalized PL Intensity
Wavelength (nm)
1100
SK 1105D00
ST 1080D15
DWELL 1085D15
SK 1095D15
ST 1125D20
DWELL 1150D15
WL (HH, LH) GaAs
2
e
500
1000
1500
2000
Delay Time(ps)
TR-PL
ALE dot: longer lifetime, thinner wetting layer
1.1
1.2
1.3
1.4
Photon Energy (eV)
PLE
1.5
III. QD LD - Modulation doping
P+ GaAs 200 nm
Contact Layer
P- GaAs 200 nm
P- Al0.70Ga0.30As 1.5 m
P-Al0.15Ga0.85As 40 nm
GaAs 50 nm
X3
P- GaAs 3 nm
GaAs 12 nm
InAs DWELL
GaAs 65 nm
n-Al0.15Ga0.85As 40 nm
n- Al0.70Ga0.30As 1.5 m
n- GaAs 200 nm
n+ GaAs Substrate
Cladding Layer
Waveguide
Layer
Al0.7GaAs
GaAs
N+
Contact Layer
InAs QD
P+
Hi T. grown spacer layer
HGTSL (High Growth Temperature Spacer Layer)1
o
o
P+ GaAs 200 nm
sub
Contact Layer
P- GaAs 200 nm
P- Al0.70Ga0.30As 1.5 m
P-Al0.15Ga0.85As 40 nm
GaAs 50 nm
X3
P- GaAs 3 nm
GaAs 12 nm
InAs DWELL
Cladding Layer
Waveguide
Layer
GaAs 65 nm
n-Al0.15Ga0.85As 40 nm
n- Al0.70Ga0.30As 1.5 m
n- GaAs 200 nm
n+ GaAs Substrate
Contact Layer
1
H.Y. Liu et al., JAP, 96, 1988, (2004) [Sheffield]
QDLD PL Spectrum
0.12
PL Intensity [a. u.]
0.10
G.S.
1302nm
(952 meV)
QDLD #1612
RT-PL
0.08
FWHM= ~32 meV
0.06
1st E.S.
1206 nm
(1028 meV)
0.04
0.02
76 meV
0.00
1000
1050
1100
1150
1200
1250
1300
1350
1400
Wave length [nm]
Large Energy separation : 76 meV
improvement in the T sensitivity1
Narrow FWHM : 32 meV
uniform size distribution of QDs
1
O.B. Shchekin et al., APL, 77, 466, (2000) [Deppe]
QDLD lasing cgaract. I
QDLD
30
L=1200 µm
L=2000 µm
W=15 µm, 1.1xIth
o
@25 C, 0.1% pulse
20 As cleaved
Optical Power [mW]
Optical Power [mW]
30
10
0
W=50 µm, 1.1xIth
o
@25 C, 0.1% pulse
20 One-sided HR
L=500 µm
100
200
300
Current [mA]
<<L-I
L-Icharacteristics
characteristicsof
of L=
L=1200
1200µm
µm&&
L=
L=2000
2000µm
µmAs-cleaved
As-cleaved QDLD>
QDLD>
L=700 µm
10
L=1000 µm
L=1500 µm
0
0
QDLD
0
100
200
Current [mA]
<<L-I
L-Icharacteristics
characteristicsof
ofL=
L=500
500
~~1500
µm
HR-coated
QDLD.
1500 µm HR-coated QDLD.>>
AS-cleaved QDLD :Jth= 493, 155 [A/cm2]
HR-coated QDLD : Jth= 634, 452, 163, 95 [A/cm2]
30
Lasing wavelength [nm]
QDLD lasing charact. II
1300
G.S.
~ 750 µµm
m
QDLD
o
@25 C, 0.1% pulse
1.1xIth
AS-cleadved
HR-coated
E.S.
1200
1000
2000
3000
Cavity length [µm]
<<Characteristics
Characteristicsof
oflasing
lasingλλvs.
vs.cavity
cavitylength
lengthLL>>
QDLD lasing charact. III
Optical Power [mW]
QDLD
W=15 µm, 1.1xIth
o
@25 C, 0.1% pulse
As cleaved
15
10
(ii)
5
0
(i)
L=1200 µm
0
100
200
Current [mA]
Normalized Intensuty [a. u.]
1.0
300
(i)
0.5
RT-PL
0.0
1.0
1302 nm
(952 meV)
(ii)
0.5
0.0
1206 nm
(1028 meV)
1200
RT-PL
1300
Wavelength [nm]
<<L-I
L-Icharacteristics
characteristicsof
ofL=
L=2000
2000µm.
µm.InInpoint
point(i),
(i),only
onlyshows
showsG.S.
G.S.lasing
lasingbut
butinin
point
point(ii),
(ii),lasing
lasingspectrum
spectrumshows
showsboth
bothG.S.
G.S.&&E.S.
E.S.lasing>
lasing>
QDLD lasing charact. IV
43
40
1325
35
2
As cleaved, LxW= 2000 x 15 µm
2
One-sided HR, LxW= 1500 x 50 µm
dλ /dT = 0.53 nm/K
1300
Optical Power [mW]
Lasing Wavelength [nm]
1350
#1612 RWG-HR
2
LxW=1500x50um
T= var., 0.1 % pulse
37
32
Jth= 89 A/cm
2
o
5 C
o
15 C
o
25 C
o
35 C
o
45 C
o
55 C
o
65 C
29
27
24
21
19
16
13
11
8
5
3
0
0
50
100
150
200
250
300
350
Current [mA]
1275
1250 QDLD
1.1xIth
o
@25 C, 0.1% pulse
1225
0
10
20
30
40
50
o
Temperature [ C]
60
70
T < 50 oC, λ shifted to
longer wavelength
due to thermally induced
band-gap narrowing
T > 55 oC, λ switched to
E.S. state
HR-coated QDLD operated
over 65 oC
QDLD lasing charart.V
5
2
As cleaved, LxW= 2000 x 15 µm (QDLD-A)
2
One-sided HR, LxW= 1500 x 50 µm (QDLD-B)
ln(Ith[mA])
To= 55 K
To= 113 K
To= 19 K
4
To= 108 K
0
10
20
30
40
50
o
Temperature [ C]
60
70
Temperature sensitivity
1. Lower growth T of
Al0.7Ga0.3As cladding layer
2. non-optimized
p-modulation doping
참고 자료
HGTSL
Sheffield G. λ= 1307 nm @RT
Jth = 32.5 A/cm2, To= 79 K
3-stack, WxL = 20x5000 µm2
KIST
λ= 1310 nm @RT
Jth= 155 A/cm2, To= 113 K
3-stack, WxL= 50x1500 µm2
P-doping
Deppe G. λ= 1290 nm @RT
Jth= 174 A/cm2, To= 200 K
5-stack, WxL= 5x650 µm2
Bhattacharya
G. λ=
Jth = 180 A/c 2 1240 nm @RT
m , To= infin
ite
5-stack, WxL
= 5x800 µm 2
Bimberg G
. λ=
Jth = 147 A 12280 nm @RT
/c
10-stack, W m , To= 150 K
xL = 100x1
500 µm 2
IV. SLD
Applications : Low coherent broadband light source for OCT
WDM-PON
-40
1.0
0.6
Current [A]
6.5
7.0
7.5
8.0
-50
Intensity [dB]
Power [W]
0.8
QD-SLD
L=2mm
W=100 µm
o
T=5 C
Pulse Operation
0.4
0.2
-60
93nm
-70
0.0
0
2
4
6
Current [A]
Max. Power ~ 0.9 W
8
-80
950
1000
1050
1100
1150
Wavelength [nm]
Max. spectral bandwidth ~ 93 nm
SLD epi structure
‰ Chirped-QD (Quantum Dot) 구조
NSC-1
AlGaAs
GaAs
AlGaAs
NSC-2
GaAs
InAs QD (1.3 µm) InAs QD (1.2 µm)
AlGaAs
NSC-3
PL intensity (arb. unit)
InAs QD (1.3 µm)
GaAs
InAs QD (1.2 µm)
InAs QD (1.3 µm)InAs QD (1.25 µm)
RT
Optical power: 15 mW
NSC-3
NSC-2
NSC-1
1000
1100
1200
1300
1400
1500
Wavelength (nm)
※NSC (NL Semiconductor)
국가지정연구실 (NRL)
SLD tapered structure
‰ Chirped-QD (Quantum Dot) SLD 제작
w
R
l
θ θ
R
g
θ
mirror
ncore
t
ncladding
Rg = Rg (λ, θ, ∆n)
∆n = ncore-ncladding
국가지정연구실 (NRL)
SLD 1
‰ QD (NSC-1) SLD
Excited state
Ground state
0
1194
NSC-1 J shape SLD
CW operating
4 CL:Ridge : 1 mm J : 1.5 mm
o
T : 20 C
1285
1275
Voltage [V]
-20
1000 mA
-30
800 mA
-40
1204
400 mA
4
2
1
50 mA
0
0
-60
1100
6
3
2
200 mA
100 mA
-50
8
Power [mW]
Power [dBm]
-10
10
5
NSC-1 SLD
Ridge : 1mm
Benting : 1.5mm
T : 20
CW operating
1200
1300
Wavelength [nm]
1400
0
200
400
600
800
1000
Current [mA]
국가지정연구실 (NRL)
SLD 2
‰ QD (NSC-2) SLD
1.2QD gr
1.3QD 1st
1.3QD gr
-30
Power [dBm]
NSC-2 SLD
Ridge : 1mm
Benting : 1.5mm
T : 20
-40 CW operating 12.
1. 50 mA
2. 100 mA
3. 200 mA
4. 300 mA
5. 400 mA
6. 500 mA
7. 600 mA
8. 700 mA
9. 800 mA
10. 900 mA
11. 1000 mA
12. 1100 mA
76 nm
74 nm
11.
10.
-50
-60
-70
1000
9.
8.
7.
6. 5. 4.
3
1100
2.
1.
1200
Wavelength [nm]
1300
1400
5
Optical power (mW)
1.2QD 1st
Jpn. J. Appl. Phys. 2005. 8
4
NSC-2 QD-SLD(J-shape)
0
- Temp : 20 C, CW
- length : 2mm
3
2
1
0
0.0
0.2
0.4
0.6
0.8
1.0
1.2
Current (A)
국가지정연구실 (NRL)
SLD 3
‰ QD (NSC-3) SLD
Appl. Phys Lett : submitted
1.25Q-1st
0
3.0
-20
1.2Q-1st
-30
1.2Q-gr
-40
Voltage [V]
98 nm
NSC-3 J shape SLD
CW operating
2.5
CL:Ridge : 1 mm J : 1.5mm
o
T : 20 C
2.0
-60
1000
1100
1.3Q-1st
1.3Q-gr
1200
1300
Wavelength (nm)
30
20
1.5
1.0
10
1.25Q-gr
-50
40
Power [mW]
Power (dB)
-10
NSC-3 J-shape SLD
- temp : 25 C
- length : 2mm
- current
200 mA
400 mA
600 mA
800 mA
0.5
0
0.0
1400 -100
0
100 200 300 400 500 600 700 800
Current [mA]
국가지정연구실 (NRL)
V. QDIP - IR PD Applications
Introduction (Types of IR detector)
Thermal
detector
Type
Disadv
Adv
Photon detector
Interband
- Bolometer type
- MCT(HgCdTe)
-Ferroelectric type
- InSb
- PN junction
- low D*
- very slow
- R.T. operation
Intersubband
- Si Schottky
- Si QWIP / QDIP
- GaAs QWIP / QDIP
- low T opeation
( <77K)
- High T. opeation(>77K)
- non uniform process
- low D*
- high D*
- fast response
- uniform and reliable
- 2D FPA (easy fab)
- fast response
Quantum Dot for IR detector
Co
n
hν
Bound-to-Quasibound
Bound-to-Continuum
DOS
Bound-to-Bound
tin
u
um
ba
nd
∆E ~ 100 meV
~ hν > kT
∆E
WL
Energy
Bound-to-Continuum Intersubband Transition
Quantum Dot instead of Quantum Well
‰ Normal incidence operation - Pyramidal shape of QD
‰ Room-temp. operation - Strong confinement by 3-D QDs
‰ Small thermionic dark current - Atomic discrete energy level
‰ Large photoconductive gain - Lack of optical phonon scattering
Issues in QDIPs
1. Operation Wavelength (3-5um, 8-12 um, > 12um)
: Bandgap engineering by precise growth, post-growth technique
2. Device performance : Operation T, Responsitivity, Detectivity
R = (e/hc)ηλG ; Responsitivity
D* = A½ R/in
; Detectivity, in ; Noise current
For high D* and high temperature operation
Low noise(dark) current, high gain are required
QD : 3-dimensional confinement of electrons
-> promising for high temperature operation of IR detector
But grown at low temperature under strained material condition
-> defects and/or misfit dislocation is generated
-> control of defects is crucial for high quality QDIP
For low dark current;
High quality material (growth)
Post-growth treatment (Thermal treatment, H-passivation)
QDIP의 제작
cleaning
Mesa etching
Metallization
Packaging
(1) Mesa etching
2 mm
Metal pattern
(0.5 mm)
(2) Metallization and RTA
Metal pattern
(0.5 mm)
2 mm
(3) Fabricated QDIP
(4) Packaging
QD IP
Quantum Dot instead of Quantum Well
2 mm
Metal pattern
(0.5 mm)
2 mm
Metal pattern
(0.5 mm)
Photocurrent (Arb. Units)
‰ Normal incidence operation - Pyramidal shape of QD
‰ Room-temp. operation - Strong confinement by 3-D QDs
18
60
140
190
1E-3
1E-4
1E-5
2000
4000
6000
8000 10000
Wavelength (nm)
K
K
K
K
Photo-Voltaic QDIP
n+ GaAs (140 nm)
22.5
QDs GaAs (18 nm)
GaAs (18 nm)
GaAs (18 nm)
Photocurrent (a.u.)
GaAs (18 nm)
Vb= 0 V
T=10 K
T=20 K
T=30 K
T=40 K
T=50 K
0.15
n+ GaAs (350 nm)
SI (100) GaAs substrate
0.10
0.05
0.00
0
100
200
300
400
Photon energy (meV)
Direct Si doping ~
2.0x1018
500
600
T=10 K
17.5
15.0
12.5
10.0
-40
-20
0
20
40
Bias voltage (mV)
cm-3
e
Bottom contact
20.0
Responsivity (mA/W)
Al0.3Ga0.7As (28 nm)
3 stacked-In 0.5 G a 0.5 As/G aAs
Q D structure
Al 0.3 Ga 0.7 As
Blocking layer
Top contact
• Photo-voltaic effect due to asymmetric band structure
induced by direct Si doping, AlGaAs barrier and
asymmetric QD shape
•
R = 21 mA/W at zero bias at λ∼ 6.2 µm (200 meV)
• Signal detected up to 50 K
• No bias -> small dark current -> high detectivity
responsivity is 70 times larger than reported value
(MNE2004, Microelectron. Eng., accepted, 2005)
Post-growth treatment(Thermal annealing)
•Purpose
- detection wavelength tuning with post-growth technique
• 300 nm SiO2 capping and annealing (T=700 0C , 1 min )
as-grown sample
thermally treated sample
T = 30 K
PL Intensity (a. u.)
0.3
n+ GaAs (140 nm)
Al0.3Ga0.7As (28 nm)
GaAs (18 nm)
QDs GaAs (18 nm)
1.109 eV
1.157 eV
0.2
0.1
GaAs (18 nm)
GaAs (18 nm)
0.0
n+ GaAs (350 nm)
SI (100) GaAs substrate
Direct Si doping ~
2.0x1018
0.9
1.0
1.1
1.2
1.3
1.4
Photon energy (eV)
cm-3
- PL peak position : blue-shifted by 48 meV.
(due to intermixing of In/Ga composition)
- PL FWHM : 60 Æ110 meV
- PL intensity : decreased
Post-growth treatment(Thermal annealing)
Dark current
Photo-current
0.10
0.01
Photo current (a. u.)
Dark current (A)
0.08
1E-3
1E-4
T = 10 K
as-grown sample
thermally treated
sample
1E-5
-0.4
-0.2
0.0
0.2
0.4
Bias Voltage (V)
178 meV
0.06
Vb=-20 mV, T=10 K
200 meV
as-grown QDIP
thermally treated
QDIP
x 10
0.04
0.02
0.00
0
100
200
300
400
500
Photon energy (meV)
-Peak position(λ∼ 6.9 µm) : red-shifted by 22 meV.
- Dark current level : 2 times larger.
⇒ due to intermixing In/Ga composition.
(due to defects generation )
- Relatively small increase than reported data*.
( might be due to SiO2 dielectric capping layer)
-Responsivity : decreased (13.2 Æ 6.2 mA/W).
⇒ due to increase of defects.
1st demonstration of detection wavelength tuning !!
*K. Stewart, et al, J. Appl. Phys. 94, 5283 (2003).
(JJAP, accepted, 2005)
VI. Summary
-
QD growth (ALE) more uniform and
thinner wetting layer
QD LD, Ith, T0
QD SLD, wide bandwidth w/ chirping
structure
QD IP, hi T operation, photovotaic
structure, post growth treatment
VII. Prospects
Future Collaboration in:
- Optical Commun devices
- Intersubband devices, far-IR, THz
- Sensors
- LED
- Bio-photonics