Optical Spectroscopy of Carbon Nanotube p-n Diodes

Optical Spectroscopy of
Carbon Nanotube p-n Junction Diodes
Ji Ung Lee
College of Nanoscale Science and Engineering
University at Albany-SUNY
p
n
6th US-Korea Forum on Nanotechnology
April 28-29, 2009
1
[email protected]
The College of Nanoscale Science & Engineering and
Albany NanoTech Complex at the University at Albany
2
[email protected]
State-of-the-Art Infrastructure
NanoFab 300S
NanoFab 300N
$50M, 150K ft2
32K Cleanroom
Completed: 03/04
$175M, 228K ft2
60K Cleanroom
Completion: 10/08
NanoFab 300E
NanoFab 200
$100M, 250K ft2
Completion: 1Q/09
$16.5M, 70K ft2
4K Cleanroom
Completed: 06/97
750K ft2 cutting-edge facilities (96,000 ft2 300mm Wafer Cleanrooms).
$4.5B investments and 2500 R&D jobs on site.
3
[email protected]
300 mm Wafer Processing Capability
ANT/CNSE will house over 125 state-of-the-art 300mm
wafer tools when build out is completed.
Designed for 32nm node & beyond but compatible with
previous generations.
• Unit process, module integration, and full flow
capability.
• Facility will have a 45nm baseline process for use
by partners.
Facility capable of 25 integrated wafer starts (WSD)
per day.
• 24/7 operation, wafer release 6 Days / Week
4
[email protected]
Device fabrication on 300mm wafers
>1000 devices/die
~100 nm features
Advanced processes
70nm
5
[email protected]
Why study the p-n diode:
• The p-n junction diode is the most fundamental of
all the semiconductor devices – it is the basis for the
majority of solid state devices.
• For fundamental understanding of semiconductors:
Example: Hall-Shockley-Read Theory.
For any new semiconductor, a proper characterization
of the p-n diode is important.
6
[email protected]
Interplay between transport and optical
properties:
• SWNT Diode Fabrication and DC Characteristics
• Optical Properties:
Photovoltaic Effect
Enhanced Optical Absorption - Excitons
• Origin of the Ideal Diode Behavior (BGR-Bandgap
Shrinkage)
7
[email protected]
Bulk p-n junction diode basics:
I
V
N-type(electrons) P-type(holes)
EC
Equilibrium
EF
EV
EC
Forward Bias
(Recombination)
8
I=Io(eqV/nKT-1)
I
EV
2
Reverse Bias
(Generation)
Diode Equation:
(ideal if n=1)
1
EC
3
V
EV
[email protected]
Electrostatic doping:
Carrier
Concentration
p
n
Split gates VG1,2
9
J.U. Lee et. al., APL: July 5, 2004
[email protected]
S
D
2 gate device
VG1
20µm
VG2
3 and 4
gate devices
10
[email protected]
CNT diode/rectifier:
(p-n or n-p diode devices)
1 10 -6
p
S
5 10
-10V
-7
D
-10V
0 10 0
-5 10 -7
S
-1 10 -6
-1.5
p
-10V
-1
n
S
D
+10V
+10V
-0.5
0
V (Volts)
DS
11
n
0.5
p
D
-10V
1
1.5
J.U. Lee et. al., APL: July 5, 2004
[email protected]
Nearly Ideal Diode Characteristics
with n~1 (1.2)
10 -7
10
-8
10
-9
p
n
p
I = I o (e
n
10 -10
qV
nK BT
− 1)
VGS1,2=+/-10V
Fit
10 -11
-0.4
-0.2
0
V
12
DS
0.2
0.4
(Volts)
[email protected]
Series Resistance Limits Current:
10 -7
Rs
10
-8
10
-9
Rs:
measured
from the
resistive
mode – due
to n-type to
metal
contact
resistance.
10 -10
10 -11
-0.4
-0.2
0
V
13
DS
0.2
0.4
(Volts)
[email protected]
Suspended SWNT Diodes:
(a)
p
n
(b)
1 µm
Suspended tube formed
based on a self-registering
technique
14
[email protected]
Ideal Diodes with Ideality Factor n=1.0 for Suspended Diodes
10-7
1.E-07
1.E-08
10-8
1.E-09
1.E-10
Fit
Data
IDS (Amps)
1.E-11
1.E-12
10-9
1.E-13
-0.5
0
0.5
10-10
Rs
n=1.0
SWNTs are
perfect,
substrates
are not.
-11
10
10-12
10-13
-0.5
15
0
VDS(V)
0.5
1
J.U. Lee, Appl. Phys. Lett. 87, 073101 (2005)
[email protected]
Photovoltaic Effect
8x10
-12
(λ =1.5 µm)
IDS (Amps)
4x10
16
-12
n
p
LED
0
Voc and Isc:
Completely define
PV properties for an
ideal diode
Voc
-4x10
-12
-8x10
-12
PV
PD
-0.2
Isc
Increase Intensity
-0.1
VDS (V)
0
0.1
J.U. Lee, Appl. Phys. Lett. 87, 073101 (2005)
[email protected]
Exciton Peaks in the Photocurrent Spectra
4x10
IDS (A)
(similar to SWNTs in solution)
-14
3
1
3x10
-14
10
-11
10
-12
10
-13
10
-14
10
-15
-0.10
-0.05
0.00
0.05
0.10
VDS(V)
ISC (A)
2
2x10
4
-14
5
3
1x10
-14
0
0.5
1
1.0
1.5
Energy (eV)
J.U. Lee et.al., Appl. Phys. Lett. 90, 053103 (2007)
17
[email protected]
E
3D
Bulk Semiconductor
18
D. O. S.
D. O. S.
D. O. S.
D. O. S.
DOS: One Electron Model
E
2D
Quantum Well
E
E
1D
Quantum Wire
0D
Quantum Dot
[email protected]
EXCITONS IN CARBON NANOTUBES
continuum
Exciton Hydrogenic Levels
n=1,2,3…
Electron-Hole Coulomb Interaction
e2
Heh = −
ε | re−rh |
results in the electron-hole binding
that forms the exciton states below
the conduction subband edge
19
[email protected]
Sommerfeld Factor: Coulomb Interaction
3D:
Excitons
Absorption
E
2D:
1D:
Coulomb
Effects
Energy
Absorption
Coulomb
Effects
Energy
Absorption
Coulomb
Effects
Energy
20
[email protected]
Sommerfeld Factor in 1D -> 0 at Eg
T. Ogawa and T. Takaghara, Phys. Rev. B 43, 14325 (1991)
21
[email protected]
Spectra with similar first energies
1 = E11
ISC (Normalized)
2.0
3 = E22
EB
2
1.5
Lack of any
features at Eg
due to
Sommerfeld
factor <1
1.0
0.5
0.6
0.8
1.0
Energy (eV)
1.2
1.4
Side bands
measure dark
exciton
J.U. Lee et.al., Appl. Phys. Lett. 90, 053103 (2007)
22
[email protected]
Comparison to Photoluminescent Data:
2.0
Intensity (a.u.)
1.6
Energy (eV)
+: Emperical Kataura
Weisman et.al. Nano Lett.
3, 1235 (2003)
100
300
-1
Raman frequency (cm )
1.2
Continuum:
1.55eV/nm
0.8
0.4
1.0
E11: 1.01eV/nm
1.2
1.4
1.6
Diameter (nm)
23
200
- E11 and E22
– Exciton-phonon
▲ - Quasipaticle Bandgap
1.8
2.0
EB: 0.54 eV/nm
[email protected]
-8
10
-9
10
-10
10
-11
10
-12
10
-13
10
-14
10
-15
-0.10
Ideal Diodes:
n=1.0
1.0
-0.05
0.00
0.05
0.10
0.15
0.20
Two mechanism for n=1.0:
1) Direct Band-to-Band
2) Diffusion of Minority Carriers
from the doped regions
3 = E22
1 = E11
30
0.9
VDS (V)
24
40
2
ISC (fA)
10
E11 (eV)
IDS (A)
Origin of the Ideal Diode Behavior and Exciton
Dissociation:
0.8
5 = E33
10
0
0.5
0.7
4
20
1.0
1.5
Energy (eV)
E11=Ea
0.6
0.5
0.4
0.5
0.6
Ea(eV)
Ea < E11 ??
[email protected]
Many-Body Renormalization of Band structure
(BGR – band gap renormalization) and Proposed
Mechanism for Exciton Dissociation:
L
n
S
np
Isc
EC
EB
2
EF
Ea
D
E11
1
p
3
Ea
pn
EV
Isc
Formation of heterointerfaces along a homogenous material
J.U. Lee, Phys. Rev. B 75, 075409 (2007)
25
[email protected]
Device Ideal for Studying BGR:
Variable Doping
with VG1,2:
• Diode follows
ideal relation with
doping.
1E-8
n
1E-9
1E-10
6V
8V
11V
p
S
D
IDS (A)
SiO2
1E-11
1E-12
VG2
VG1
L
1E-13
1E-14
1E-15
-0.10
-0.05
0.00
0.05
VDS (V)
26
0.10
0.15
0.20
• Evidence of
strong BGR: Io
when Doping .
w/o BGR Io when
Doping .
[email protected]
Origin of increase in Io with Doping:
Minority
Carriers
Ef
P type
semiconductor
27
Increase
Doping
No
shrinkage
of the
band gap
Ef
w/o BGR:
minority carrier
decreases
Shrinkage
of the
band gap
Ef
w/ BGR:
minority
carrier
increases!
[email protected]
Conclusions:
• Bipolar devices are more fun to study.
• How do neutral excitons dissociate to generate
large photocurrents?
• Window to the study of many-body effects:
BGR, biexctions, etc…
Funding: NSF, NRI/INDEX, IFC, IBM and
UAlbany
28
[email protected]
Future Work:
Graphene p-n junctions: Optics-like manipulation of electrons
n-type
p-type
1,2...layer
graphene
flake
Split Gates
29
[email protected]